GNKSA approval of M2
Asbjørn Ulsberg
I haven't evaluated Opera to the GNKSA evaluation form[2] yet,
but I already know it will fail a lot of those tests.
Is it a goal that M2 will have the GNKSA shield some day, or
isn't that of any interest to Opera? Just wondering. :-)
____
[1] The Good Net-Keeping Seal of Approval:
<url:http://www.newsreaders.com/gnksa/>
[2] <url:http://www.newsreaders.com/gnksa/cgi/gnksa-form.cgi>
--
Asbjørn Ulsberg -=|=- http://virtuelvis.com/quark
"He's a loathsome offensive brute, yet I can't look away"
Rijk van Geijtenbeek
wrote:
> I just wonder what Opera does to aspire GNKSA[1] approval?
> I haven't evaluated Opera to the GNKSA evaluation form[2] yet,
> but I already know it will fail a lot of those tests.
Actually, I found only very few fails, and mostly for SHOULDs, not for
MUSTs, afaict. And some fails are actually not that bad. For example:
Opera does not give a warning when trying to post an empty message (a
must), but it fails to send the message anyway (a should), giving the user
an error dialog. This could be done more gracefully (like when Opera gives
a proper warning that it will not send without a Subject), but it keeps
usenet clean. The GNKSA pages [1] seem to be abondened, or at least not
updated, for three years now BTW.
> Is it a goal that M2 will have the GNKSA shield some day, or
> isn't that of any interest to Opera? Just wondering. :-)
I know our usenet expert Frode is well aware of GNSKA. He pointed out
Opera's failures before he joined the company :)
Interesting reading:
http://groups.google.com/groups?selm=Xns900F8B545frodegillagressono%40news.eunet.no
http://groups.google.com/groups?selm=Xns9224D144ABEFDfrodegillbrekkelia3d%40193.69.113.75
Here's a list of GNKSA standards where the answer is 'No', afaict:
2c Provides clear, separate command for replying by e-mail ([2], MUST)
2d Uses standard terminology ([3] SHOULD)
3b Warns about, or prevents, posting to large numbers of groups (SHOULD)
3c Strongly encourages setting of a "Followup-To: " for large crossposts
(SHOULD)
9d When both post- and mailing, precedes mail copy with clear notification
stating so (SHOULD)
10e Offers users some means of indicating which part(s) to followup to
(SHOULD)
13b Allows superseding articles (SHOULD)
14a Articles are posted as edited, with linebreaking intact ([4] SHOULD)
14b Warns about lines over 80 characters ([4] SHOULD)
15b Warns against or refuses to use excessive signatures (SHOULD)
16a Warns when attempting to post empty articles ([5] MUST)
16c Warns when post articles containing quoted material only (MUST)
16d Refuses posting quoted-text-only articles (SHOULD)
16e Warns against posting articles multiply ([6] MUST)
16f Prevents multiple posting entirely ([6] SHOULD)
18a Allows filtering out annoying articles (killing) ([7] SHOULD)
I didn't test these:
7d Keeps as many References from original as fit
7e Does not propagate broken Message-IDs in original References
Notes:
[1] http://www.newsreaders.com/gnksa/cgi/gnksa-form.cgi
[2] Instead you have to do 'Reply to all' and then trim the newsgroup
entry. Or you use the Quick Reply field (horror).
[3] The word 'followup' is not used, but 'reply' instead
[4] Opera always uses format=flowed, which obscures the issue being
handled with standard 14 "Try to respect the 80-character line-length
conventions". By default (but user controllable) a 76 character linelength
is enforced, but the compose window does not show this. More on f=f:
http://www.joeclark.org/ffaq.html
[5] Opera doesn't send these messages, so the end-result is fine, if a bit
user-unfriendly. The situation isn't likely to occur by accident, with the
default signature. GNSKA doesn't say whether sending only a few linefeeds
and a signature should be prevented.
[6] I guess this is possible in Opera by opening messages from the Sent
folder again.
[7] You can quite easily create (and add to) an 'ignore' filter that marks
certain messages as read, but you can't for example refuse messages that
are crossposted to more than x groups.
--
The Web is a procrastination apparatus: | Rijk van Geijtenbeek
It can absorb as much time as | Documentation & QA
is required to ensure that you | Opera Software ASA
won't get any real work done. - J.Nielsen | mailto:ri...@opera.com N
Asbjørn Ulsberg
>> I just wonder what Opera does to aspire GNKSA[1] approval?
>
> Actually, I found only very few fails, and mostly for SHOULDs,
> not for MUSTs, afaict.
I haven't walket through the form _very_ closely, but am planning
to do the full evaluation some day.
> And some fails are actually not that bad.
No, it's not as horrific as Outlook Express (at least was), but
there's no good reason why Opera shouldn't pass the GNKSA validation
with a bright and shining star.
> The GNKSA pages [1] seem to be abondened, or at least not updated,
> for three years now BTW.
NNTP hasn't evolved in a lot more years than that. :-) The GNKSA are
pretty static rules, made to fit the medium NNTP as is. As long as
the NNTP standard doesn't evolve, there's no apparent reason why the
GNKSA should either, no?
> I know our usenet expert Frode is well aware of GNSKA. He pointed out
> Opera's failures before he joined the company :)
Goodie. I know he used Xnews earlier and are aware of stuff like this,
but I feel that he has sunken into some kind of void inside Opera,
and that he doesn't get the time he needs to make Opera the best news-
reader ever. I'm sure that's what he'd do, if he had the time.
> 2c Provides clear, separate command for replying by e-mail ([2], MUST)
Yes, this is critical. It's terrible to give users the idea that NNTP
is the same protocol as POP3 and IMAP, though the messages transfered
over the protocols is pretty much the same.
> 3b Warns about, or prevents, posting to large numbers of groups (SHOULD)
I haven't tested this, but I have the opportunity to do it at work where
I have Hamster innstalled. Do you know for sure that Opera doesn't warn
about this?
> 3c Strongly encourages setting of a "Followup-To: " for large crossposts
> (SHOULD)
Same as above.
> 13b Allows superseding articles (SHOULD)
I never do this myself, so not an important issue to me, but a SHOULD is
a SHOULD is a SHOULD. :-)
> 14a Articles are posted as edited, with linebreaking intact ([4] SHOULD)
I feel Opera is breaking this a bit, yes.
> 14b Warns about lines over 80 characters ([4] SHOULD)
This is a really irritating one. Xnews has a hard break on 80 (or what
you define) characters, and if you type more than that, the text is
wrapped to a new line. Also, Xnews has an ugly-looking ruler at the
top of the edit-area to show where the break is going to occur.
> 16f Prevents multiple posting entirely ([6] SHOULD)
Ah, wouldn't this be great.
> 18a Allows filtering out annoying articles (killing) ([7] SHOULD)
You can kind of do this with filters, but it's not the same. Both
scoring and filtering is something I miss deeply in M2 (which I had
in Xnews).
> 7d Keeps as many References from original as fit
I've tested it, and Opera does this quite well. It makes sure that
the References-header doesn't exceed 998 characters, and starts
cutting it down when the header reaches the 998-mark.
It also seems as it keeps the first message-id at all time, and
starts cutting at message-id number two, so it always has the
start and end reference of a thread. This is good behavior.
> 7e Does not propagate broken Message-IDs in original References
Haven't tested this yet, but will at work, where I have Xnews
where I can fake a message-id.
> [2] Instead you have to do 'Reply to all' and then trim the newsgroup
> entry. Or you use the Quick Reply field (horror).
Blech. This is terrible behavior.
> [3] The word 'followup' is not used, but 'reply' instead
Yup, not good.
> [4] Opera always uses format=flowed, which obscures the issue being
> handled with standard 14 "Try to respect the 80-character line-length
> conventions". By default (but user controllable) a 76 character
> linelength is enforced, but the compose window does not show this.
Yup. Not good at all. I hate it, infact.
> [5] Opera doesn't send these messages, so the end-result is fine, if a
> bit user-unfriendly. The situation isn't likely to occur by accident,
> with the default signature.
I don't think it's any better that Opera allows users to post messages
with nothing but a signature in it, but I guess most newsreaders allow
this. It's utterly simple to test against, thought;
if (body.trim() == signature) throw new EmptyBodyException();
> GNSKA doesn't say whether sending only a few linefeeds and a signature
> should be prevented.
No, but common sense kind of says it should, no? :-)
> [6] I guess this is possible in Opera by opening messages from the Sent
> folder again.
Opera should have a strict index over what articles are posted any
way, to do automatic scoring of messages and threads an author have
written or participated in.
> [7] You can quite easily create (and add to) an 'ignore' filter that
> marks certain messages as read, but you can't for example refuse
> messages that are crossposted to more than x groups.
True. Scoring and killing is a missed feature.
Rijk van Geijtenbeek
wrote:
> On Mon, 15 Mar 2004 02:16:12 +0100, Rijk van Geijtenbeek wrote:
>
>>> I just wonder what Opera does to aspire GNKSA[1] approval?
>>
>> Actually, I found only very few fails, and mostly for SHOULDs,
>> not for MUSTs, afaict.
>
> I haven't walket through the form _very_ closely, but am planning
> to do the full evaluation some day.
It didn't take me that long yesterday :)
>> And some fails are actually not that bad.
>
> No, it's not as horrific as Outlook Express (at least was), but
> there's no good reason why Opera shouldn't pass the GNKSA validation
> with a bright and shining star.
>
>> The GNKSA pages [1] seem to be abondened, or at least not updated,
>> for three years now BTW.
>
> NNTP hasn't evolved in a lot more years than that. :-) The GNKSA are
> pretty static rules, made to fit the medium NNTP as is. As long as
> the NNTP standard doesn't evolve, there's no apparent reason why the
> GNKSA should either, no?
The lack of mention of Unicode/UTF8 support and format=flowed is a pity. A
program like Gravity gets a seal, while it is (I've heard) unsuited for
handling UTF-8. People keep asking for hacks to send all kinds of
encodings, because so many programs still in use can't handle UTF-8. And
people keep using clients that don't do format=flowed display, which would
make their experience so much nicer, and would make the requests for
(automatic or manual) reformatting quoted material irrelevant. So, I think
GNKSA 2.1 MUST demand UTF-8 support and SHOULD encourage format=flowed
support.
The necoiding issue is mentioned on the 'todo' page (unchanged since early
2001). Some submissions are listed as being worked on (also since early
2001).
>> I know our usenet expert Frode is well aware of GNSKA. He pointed out
>> Opera's failures before he joined the company :)
>
> Goodie. I know he used Xnews earlier and are aware of stuff like this,
> but I feel that he has sunken into some kind of void inside Opera,
> and that he doesn't get the time he needs to make Opera the best news-
> reader ever. I'm sure that's what he'd do, if he had the time.
I think so as well. Let's get out of this IMAP and SMTP business and
concentrate on the really important protocol: NNTP :)
>> 2c Provides clear, separate command for replying by e-mail ([2], MUST)
>
> Yes, this is critical. It's terrible to give users the idea that NNTP
> is the same protocol as POP3 and IMAP, though the messages transfered
> over the protocols is pretty much the same.
I don't think it is so terrible, and at this moment many people don't use
their real email address in postings. Trying to reply by mail only is not
the most important feature IMHO.
>> 3b Warns about, or prevents, posting to large numbers of groups (SHOULD)
>
> I haven't tested this, but I have the opportunity to do it at work where
> I have Hamster innstalled. Do you know for sure that Opera doesn't warn
> about this?
No, not for sure. But I've never heard about it, and I haven't seen the
error message in the language file for example.
>> 3c Strongly encourages setting of a "Followup-To: " for large
>> crossposts (SHOULD)
>
> Same as above.
I'm quite sure there is no UI for that encouragement.
>> 13b Allows superseding articles (SHOULD)
>
> I never do this myself, so not an important issue to me, but a SHOULD is
> a SHOULD is a SHOULD. :-)
But a SHOULD is not a MUST :)
>> 14a Articles are posted as edited, with linebreaking intact ([4] SHOULD)
>
> I feel Opera is breaking this a bit, yes.
>
>> 14b Warns about lines over 80 characters ([4] SHOULD)
>
> This is a really irritating one. Xnews has a hard break on 80 (or what
> you define) characters, and if you type more than that, the text is
> wrapped to a new line. Also, Xnews has an ugly-looking ruler at the
> top of the edit-area to show where the break is going to occur.
How does Xnews handle large URLs (sending and receiving)? I find that, in
practice, Opera behaves very well, as long as people don't get confused
and start entering their own ENTERs...
>> 16f Prevents multiple posting entirely ([6] SHOULD)
>
> Ah, wouldn't this be great.
>
>> 18a Allows filtering out annoying articles (killing) ([7] SHOULD)
>
> You can kind of do this with filters, but it's not the same. Both
> scoring and filtering is something I miss deeply in M2 (which I had
> in Xnews).
I've never figured out how to use Xnews properly in this regard. But some
simple 'watch' and 'hide' functionality for both threads and persons is
nice. We've got some of it, but not really enough.
>> 7d Keeps as many References from original as fit
>
> I've tested it, and Opera does this quite well. It makes sure that
> the References-header doesn't exceed 998 characters, and starts
> cutting it down when the header reaches the 998-mark.
>
> It also seems as it keeps the first message-id at all time, and
> starts cutting at message-id number two, so it always has the
> start and end reference of a thread. This is good behavior.
>
>> 7e Does not propagate broken Message-IDs in original References
>
> Haven't tested this yet, but will at work, where I have Xnews
> where I can fake a message-id.
OK.
>> [2] Instead you have to do 'Reply to all' and then trim the newsgroup
>> entry. Or you use the Quick Reply field (horror).
>
> Blech. This is terrible behavior.
See above - how often do you reply by mail only? And without having to
manually edit the address anyway?
>> [3] The word 'followup' is not used, but 'reply' instead
>
> Yup, not good.
The perils of the integrated interface. I doubt this will be changed
anytime soon.
>> [4] Opera always uses format=flowed, which obscures the issue being
>> handled with standard 14 "Try to respect the 80-character line-length
>> conventions". By default (but user controllable) a 76 character
>> linelength is enforced, but the compose window does not show this.
>
> Yup. Not good at all. I hate it, infact.
What exactly? Format=flowed in general, or that the compose window doesn't
indicate where it will break? With format=flowed, it will not matter when
your readers are using capable software (Eudora, Hotmail, Mozilla Mail,
Opera, Apple Mail), because their display adapts to the window width,
which is much nicer. Unless you are composing ascii art.
>> [5] Opera doesn't send these messages, so the end-result is fine, if a
>> bit user-unfriendly. The situation isn't likely to occur by accident,
>> with the default signature.
>
> I don't think it's any better that Opera allows users to post messages
> with nothing but a signature in it, but I guess most newsreaders allow
> this. It's utterly simple to test against, thought;
>
> if (body.trim() == signature) throw new EmptyBodyException();
>
>> GNSKA doesn't say whether sending only a few linefeeds and a signature
>> should be prevented.
>
> No, but common sense kind of says it should, no? :-)
Right. People pressing 'Reply', and then 'Send' while forgetting to
actually enter your own text are also an annoyance, and it would be nice
if Opera would warn about that.
>> [6] I guess this is possible in Opera by opening messages from the Sent
>> folder again.
>
> Opera should have a strict index over what articles are posted any
> way, to do automatic scoring of messages and threads an author have
> written or participated in.
>
>> [7] You can quite easily create (and add to) an 'ignore' filter that
>> marks certain messages as read, but you can't for example refuse
>> messages that are crossposted to more than x groups.
>
> True. Scoring and killing is a missed feature.
--
Asbjørn Ulsberg
>> I haven't walket through the form _very_ closely, but am planning
>> to do the full evaluation some day.
>
> It didn't take me that long yesterday :)
Did you submit it as well?
>> As long as the NNTP standard doesn't evolve, there's no apparent
>> reason why the GNKSA should either, no?
>
> The lack of mention of Unicode/UTF8 support and format=flowed is a pity.
True.
> A program like Gravity gets a seal, while it is (I've heard) unsuited
> for handling UTF-8.
Xnews doesn't handle UTF-8 either, but still has the GNKSA seal. I think
non-UTF-8 support is OK, but newer clients SHOULD support it.
> People keep asking for hacks to send all kinds of encodings, because
> so many programs still in use can't handle UTF-8.
Yup. In Norway, the problem is kind of limited, because everyone can just
default to ISO-8859-1. But in international groups, it becomes a lot more
complicated, and the easy solution is of course UTF-8.
> And people keep using clients that don't do format=flowed display, which
> would make their experience so much nicer, and would make the requests
> for (automatic or manual) reformatting quoted material irrelevant.
I don't fully agree with that. Even with 'format=flowed', I try to make
my articles pretty, keep the lines within 80 characters, etc. I think
this is something good, at least that I _can_ control this. I don't think
I have enough control in Opera, and it makes me very anxious.
> So, I think GNKSA 2.1 MUST demand UTF-8 support and SHOULD encourage
> format=flowed support.
I agree to UTF-8, but the 'format=flowed' doesn't really mean diddlysquat
to me. But I don't mind seeing it as a part of the GNKSA.
> The necoiding issue is mentioned on the 'todo' page (unchanged since
> early 2001). Some submissions are listed as being worked on (also since
> early 2001).
Maybe we should investigate in how we can get the document updated?
> I think so as well. Let's get out of this IMAP and SMTP business and
> concentrate on the really important protocol: NNTP :)
Yea! #:-D
> I don't think it is so terrible, and at this moment many people don't
> use their real email address in postings. Trying to reply by mail only
> is not the most important feature IMHO.
My problem lies in the fact that news and e-mail isn't clearly enough
separated in M2. That's what I'm talking about.
>>> 3b Warns about, or prevents, posting to large numbers of groups
>>
>> Do you know for sure that Opera doesn't warn about this?
>
> No, not for sure. But I've never heard about it, and I haven't seen the
> error message in the language file for example.
Ok. Let's say it doesn't, then. :-)
>>> 3c Strongly encourages setting of a "Followup-To: " for large
>>> crossposts (SHOULD)
>>
>> Same as above.
>
> I'm quite sure there is no UI for that encouragement.
:-\
>> a SHOULD is a SHOULD is a SHOULD. :-)
>
> But a SHOULD is not a MUST :)
Not quite, but almost. And for a supreme piece of software like Opera,
every SHOULD ought to be upgradet to a MUST (now you might ask what
happens to MUST then? Does it go all the way around the circle and
end up as a MAY? ;).
> How does Xnews handle large URLs (sending and receiving)?
It cheats. You can select a large peice of text and right-click
'Edit URL', edit away all the nastyness and then execute the URL to
the operating system. It supports <url:..> schemes partly, but the
'Edit URL' is the base support Xnews has for URIs.
> I find that, in practice, Opera behaves very well, as long as people
> don't get confused and start entering their own ENTERs...
Hm. This should work, as long as the URI is enclosed in <url:..>
brackets.
> I've never figured out how to use Xnews properly in this regard.
The best way to do scoring in Xnews, is to edit the score file and
manually hack together a regex which matches your criterias. I've
hacked together a regex which recognizes 'Asbjørn Ulsberg' in the
'From' header.
Every post with a 'Reference' set to an article posted by me, also
gets scored up (and highlighted). Also, the whole thread gets a
higher score than usual (and thus gets prioritized higher in the
sort order).
> But some simple 'watch' and 'hide' functionality for both threads
> and persons is nice. We've got some of it, but not really enough.
I don't think we have anything compared to Xnews. Xnews is just
brilliant here, although a bit too difficult to set up. If Opera
could embrace all that's good in Xnews somehow, it would be just
perfect!
> See above - how often do you reply by mail only?
Seldom, but when I do, I would expect exactly the same as the GNKSA.
> And without having to manually edit the address anyway?
Also seldom, but removing an '.invalid' part of the e-mail address
is different (and more common) than removing all sorts of other
headers because M2 doesn't have implemented a way to directly reply
by e-mail.
>>> [3] The word 'followup' is not used, but 'reply' instead
>>
>> Yup, not good.
>
> The perils of the integrated interface. I doubt this will be changed
> anytime soon.
Me too. It's not critical either. Even if there is a technical difference
between follow-up's and reply's, I don't find it that important. 'Reply'
works fine, if people just know what they're replying to. Here's where
I think M2 is too vague.
>> Yup. Not good at all. I hate it, infact.
>
> What exactly? Format=flowed in general, or that the compose window
> doesn't indicate where it will break?
Mostly that the compose window doesn't indicate where it will break. But
I tend to think 'format=flowed' gives me way too long lines as well, and
I haven't found a good and simple way to control this in M2 yet.
> With format=flowed, it will not matter when your readers are using
> capable software (Eudora, Hotmail, Mozilla Mail, Opera, Apple Mail),
> because their display adapts to the window width, which is much nicer.
True, but it also takes away a bit of the control, I feel. It should
be optional, at least.
> Unless you are composing ascii art.
Do not underestimate the power of ASCII! I draw in ASCII all the time,
actually. Attaching images is no-no, and uploading to a web site takes
way too much time. :-)
>> No, but common sense kind of says it should, no? :-)
>
> Right. People pressing 'Reply', and then 'Send' while forgetting to
> actually enter your own text are also an annoyance, and it would be nice
> if Opera would warn about that.
Yep. Especially when it's so frikkin' easy! ;-)
Adam Bailey
"Rijk van Geijtenbeek" <ri...@opera.com.invalid> wrote:
> The GNKSA pages [1] seem to be abondened, or at least not
> updated, for three years now BTW.
The GNKSA document has not been updated (the USEFOR IETF group[1] is
currently working on new Usenet standards, to which the GNKSA will
have to adapt), but there are new evaluations from time to time. There
aren't that many new newsreaders being developed, and the existing
ones either aren't updated very often, or already pass the GNKSA.
[1] http://www.landfield.com/usefor/
--
Adam Bailey | Chicago, Illinois
ad...@lull.org | Finger/Web for PGP & S/MIME
ada...@aol.com | http://www.lull.org/adam/
Adam Bailey
Asbjørn Ulsberg <asbj...@hotmail.com> wrote:
> On Mon, 15 Mar 2004 02:16:12 +0100, Rijk van Geijtenbeek wrote:
>
> > > I just wonder what Opera does to aspire GNKSA[1] approval?
> >
> > Actually, I found only very few fails, and mostly for SHOULDs,
> > not for MUSTs, afaict.
>
> I haven't walket through the form _very_ closely, but am planning
> to do the full evaluation some day.
Let me know if you need any assistance when you do. There is a suite
of tools that can make the job easier.
> > And some fails are actually not that bad.
>
> No, it's not as horrific as Outlook Express (at least was),
Outlook Express hasn't dramatically improved its GNKSA compliance
since I did the evaluation of 5.0 in 1999.
-snip-
> > 13b Allows superseding articles (SHOULD)
>
> I never do this myself, so not an important issue to me, but a SHOULD is
> a SHOULD is a SHOULD. :-)
Maybe you'd do it if you could. :)
I do it all the time.
> > GNSKA doesn't say whether sending only a few linefeeds and a signature
> > should be prevented.
>
> No, but common sense kind of says it should, no? :-)
Yes, and I'll flag that in an evaluation. Outlook Express will
actually add blank lines to get around a news server's empty article
restriction!
Adam Bailey
Asbjørn Ulsberg <asbj...@hotmail.com> wrote:
> On Tue, 16 Mar 2004 01:30:49 +0100, Rijk van Geijtenbeek wrote:
>
>>> I haven't walket through the form _very_ closely, but am planning
>>> to do the full evaluation some day.
>>
>> It didn't take me that long yesterday :)
A proper evaluation takes me the better part of a day. That includes a
pretty extensive regiment of tests to try and "break" the newsreader.
>>> As long as the NNTP standard doesn't evolve, there's no apparent
>>> reason why the GNKSA should either, no?
>>
>> The lack of mention of Unicode/UTF8 support and format=flowed is a
>> pity.
>
> True.
>
> > A program like Gravity gets a seal, while it is (I've heard) unsuited
> > for handling UTF-8.
>
> Xnews doesn't handle UTF-8 either, but still has the GNKSA seal.
You can blame that on me. I wasn't well versed in UTF-8 when I did the
Xnews evaluation.
-snip-
>> So, I think GNKSA 2.1 MUST demand UTF-8 support and SHOULD
>> encourage format=flowed support.
>
> I agree to UTF-8, but the 'format=flowed' doesn't really mean diddlysquat
> to me. But I don't mind seeing it as a part of the GNKSA.
It'll mean something to you when you can resize a window and have
everything still look good.
>> The necoiding issue is mentioned on the 'todo' page (unchanged
>> since early 2001). Some submissions are listed as being worked on
>> (also since early 2001).
>
> Maybe we should investigate in how we can get the document updated?
Most GNKSA discussion occurs in news.software.readers. I've also
emailed a pointer to this thread to a couple of the leading GNKSA
advocates.
van Grieg
wrote:
> The lack of mention of Unicode/UTF8 support and format=flowed is a pity.
> A program like Gravity gets a seal, while it is (I've heard) unsuited
> for handling UTF-8. People keep asking for hacks to send all kinds of
> encodings, because so many programs still in use can't handle UTF-8.
Last time I checked (it was a couple of years ago though), there were no
newsreaders which could handle UTF-8, except for Outlook Express. Agent
couldn't even work with any charsets beyond 8859-1, IIRC.
On the other hand, I have yet to find a single newsgroup that would use
Unicode. So I don't think it is a _real_ problem.
--
van Grieg
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/
Rijk van Geijtenbeek
wrote:
> On Tue, 16 Mar 2004 01:30:49 +0100, Rijk van Geijtenbeek wrote:
>
>>> I haven't walket through the form _very_ closely, but am planning
>>> to do the full evaluation some day.
>>
>> It didn't take me that long yesterday :)
>
> Did you submit it as well?
Nope. I'll send you the URL to the filled in form in private mail.
..
>> And people keep using clients that don't do format=flowed display,
>> which would make their experience so much nicer, and would make the
>> requests for (automatic or manual) reformatting quoted material
>> irrelevant.
>
> I don't fully agree with that. Even with 'format=flowed', I try to make
> my articles pretty, keep the lines within 80 characters, etc. I think
> this is something good, at least that I _can_ control this. I don't think
> I have enough control in Opera, and it makes me very anxious.
Opera keeps the lines within 76 characters automatically, and shows what
it does with quoted text. I'd like a thin vertical line to indicate where
the wrapping will occur - but this is impossible to implement when you
allow proportinal fonts in the compose window. Using softwrapping in the
compose window might be what you want. But as most of my mail and news
goes to recipients who grok f=f, they'll get different linebreaks anyway
depending on their window width. Which is fine by me, just as I don't care
how wide their browser windows are when they view my HTML pages.
[..]
>> I don't think it is so terrible, and at this moment many people don't
>> use their real email address in postings. Trying to reply by mail only
>> is not the most important feature IMHO.
>
> My problem lies in the fact that news and e-mail isn't clearly enough
> separated in M2. That's what I'm talking about.
For Opera, it is a feature that mail and news are tightly integrated.
Don't expect chganges on that front. It should be rather easy to have a
separate toolbar for newsgroup windows, which would allow the use of
special terminology ('followup', 'reply-by-mail'). But I don't think we
really *want* to use different terminology.
[..]
>>> a SHOULD is a SHOULD is a SHOULD. :-)
>>
>> But a SHOULD is not a MUST :)
>
> Not quite, but almost. And for a supreme piece of software like Opera,
> every SHOULD ought to be upgradet to a MUST (now you might ask what
> happens to MUST then? Does it go all the way around the circle and
> end up as a MAY? ;).
I don't like the news client popping up warnings all the time, and
refusing to do something is even worse. But some of the shoulds and musts
are important enough, imho these should be fixed:
3b Warns about, or prevents, posting to large numbers of groups (SHOULD)
9d When both post- and mailing, precedes mail copy with clear notification
stating so (SHOULD)
10e Offers users some means of indicating which part(s) to followup to
(SHOULD)
16a Warns when attempting to post empty articles (MUST)
16c Warns when post articles containing quoted material only (MUST)
18a Allows filtering out annoying articles (killing) (SHOULD)
..
>> But some simple 'watch' and 'hide' functionality for both threads
>> and persons is nice. We've got some of it, but not really enough.
>
> I don't think we have anything compared to Xnews. Xnews is just
> brilliant here, although a bit too difficult to set up. If Opera
> could embrace all that's good in Xnews somehow, it would be just
> perfect!
If 'Active threads' catched more threads, and/or there was an option to
send a thread there, it would help a lot. Opera does show a different icon
in front of posting from people in your contacts, which helps. It would be
even better if I could see which threads contain postins from people I
know without expanding them. But I think the number of columns is large
enough already, and I rather keep an uncluttered view of the message list.
I find the display in Opera much prettier than TheBat or Agent or Eudora
or Xnews... But now I'm drifting off-topic :)
Rijk van Geijtenbeek
> On Tue, 16 Mar 2004 01:30:49 +0100, Rijk van Geijtenbeek
> <ri...@opera.com> wrote:
>
>
>> The lack of mention of Unicode/UTF8 support and format=flowed is a
>> pity. A program like Gravity gets a seal, while it is (I've heard)
>> unsuited for handling UTF-8. People keep asking for hacks to send all
>> kinds of encodings, because so many programs still in use can't handle
>> UTF-8.
>
> Last time I checked (it was a couple of years ago though), there were no
> newsreaders which could handle UTF-8, except for Outlook Express. Agent
> couldn't even work with any charsets beyond 8859-1, IIRC.
Agent (1.8 and later at least) have extensive controls to choose
encodings, though the prefs are a bit confusing.
> On the other hand, I have yet to find a single newsgroup that would use
> Unicode. So I don't think it is a _real_ problem.
:) Maybe I should enter a smiley character here... ☺
Upon entering this character, Opera warns me that it will have to send
this message as UTF-8. If I refuse, it will send the message with a ?
instead of the ☺.
Rijk van Geijtenbeek
> In article <opr4vpta...@news.opera.com>,
> "Rijk van Geijtenbeek" <ri...@opera.com.invalid> wrote:
>
>> The GNKSA pages [1] seem to be abondened, or at least not
>> updated, for three years now BTW.
>
> The GNKSA document has not been updated (the USEFOR IETF group[1] is
> currently working on new Usenet standards, to which the GNKSA will
> have to adapt),
Understandably.
> but there are new evaluations from time to time.
I now see that <URL:http://www.gnksa.org/gnksa-evaluations.html> has a
later than 2001 'Last modification' date: "Jul 27 2002 (Sep 3252 1993) by
JS." This is not entirely clear to me though :)
> There
> aren't that many new newsreaders being developed, and the existing
> ones either aren't updated very often, or already pass the GNKSA.
This page made me think that nothing had been done at all since 2001:
<URL:http://www.gnksa.org/in-progress.txt>, so I wondered if it made sense
to send an evaluation of Opera 7.
van Grieg
wrote:
>> On the other hand, I have yet to find a single newsgroup that would use
>> Unicode. So I don't think it is a _real_ problem.
>
> :) Maybe I should enter a smiley character here... ☺
> Upon entering this character, Opera warns me that it will have to send
> this message as UTF-8. If I refuse, it will send the message with a ?
> instead of the ☺.
>
<sigh>
Ironically, the very people who were supposed to benefit most from Unicode
are often the fiercest antagonists of its wide acceptance. Few languages
have suffered as much from the encoding zoo as has Russian - we have like
5 of them, although only 2 are widespread.
Unfortunately, while being a wonderful idea overall, Unicode causes a
couple of really bad problems. The first one is compatibility. It's still
much better to send a few messed up characters than a totally unreadable
message. The second one is overhead. With large mails/pages, Unicode can
double the size. On slow dialup it is a _really_ annoying thing, believe
it or not.
Asbjørn Ulsberg
> Unfortunately, while being a wonderful idea overall, Unicode causes a
> couple of really bad problems. The first one is compatibility.
As Unicode is built to be backward compatible with ASCII, I don't see
the compatibility-issues. What are they related to, really?
> It's still much better to send a few messed up characters than a totally
> unreadable message.
It's only non-ASCII characters (and not all of those, either) which are
placed differently in e.g. UTF-8, so as long as the language is written
in mostly ASCII characters, most of the message will also be readable,
even without UTF-8 support.
> The second one is overhead. With large mails/pages, Unicode can double
> the size. On slow dialup it is a _really_ annoying thing, believe it or
> not.
Bat quotation, attachments etc., caused by bad newsreaders like Outlook
Express does imho much more damage and overhead than UTF-8 ever would.
Asbjørn Ulsberg
> Last time I checked (it was a couple of years ago though), there were
> no newsreaders which could handle UTF-8, except for Outlook Express.
I think Emacs has had UTF-8 support for quite a while. Not sure about
other clients, thought.
Asbjørn Ulsberg
>> I haven't walket through the form _very_ closely, but am planning
>> to do the full evaluation some day.
>
> Let me know if you need any assistance when you do. There is a suite
> of tools that can make the job easier.
I'll let you know now. Which tools are you talking about? Feel free to
add me up on MSN Messenger, Gaim or something similar, if you have such
applications installed; asbj...@hotmail.com.
> Outlook Express hasn't dramatically improved its GNKSA compliance
> since I did the evaluation of 5.0 in 1999.
No, infact, it has aggravated. :-)
>>> 13b Allows superseding articles (SHOULD)
>
> Maybe you'd do it if you could. :)
Maybe. But isn't server support an issue here, as with cancellation,
as well? Hence, makes the functionality pretty pointless, imho.
>>> GNSKA doesn't say whether sending only a few linefeeds and a signature
>>> should be prevented.
>>
>> No, but common sense kind of says it should, no? :-)
>
> Yes, and I'll flag that in an evaluation.
Mhm.
> Outlook Express will actually add blank lines to get around a news
> server's
> empty article restriction!
OMG! It trims leading space in every other scenario (damaging the signature
separator '-- \n'), but here it adds it. Great! :-\
Asbjørn Ulsberg
>> Did you submit it as well?
>
> Nope. I'll send you the URL to the filled in form in private mail.
Yes, please do. Send it to <url:mailto:asb...@tigerstaden.no>.
> Opera keeps the lines within 76 characters automatically, and shows what
> it does with quoted text.
I don't think it shows clearly enough. I want a ruler, goddamnit! #:-D
> I'd like a thin vertical line to indicate where the wrapping will occur
> - but this is impossible to implement when you allow proportinal fonts
> in the compose window.
True, but it's just to show the ruler only if the font chosen is monospace
(whoever in their right minds will choose a proportional font for news
or e-mail?! ;-)).
> Using softwrapping in the compose window might be what you want.
Yup. Xnews does this sooo beautifully.
> But as most of my mail and news goes to recipients who grok f=f,
> they'll get different linebreaks anyway depending on their window width.
Yes, but people without f=f will have it purty.
> Which is fine by me, just as I don't care how wide their browser
> windows are when they view my HTML pages.
Goodie.
>> My problem lies in the fact that news and e-mail isn't clearly enough
>> separated in M2. That's what I'm talking about.
>
> For Opera, it is a feature that mail and news are tightly integrated.
I'm not talking about message-handling here, but protocol-handling. I
really love the fact that M2 indexes both news, e-mail and newsfeeds as
the same kinds of items (but I would like newsfeed items to have a
distinguished icon in message lists), but I dislike the fact that there's
no apparent line between e-mail and news posting etc.
> Don't expect chganges on that front. It should be rather easy to have a
> separate toolbar for newsgroup windows, which would allow the use of
> special terminology ('followup', 'reply-by-mail'). But I don't think we
> really *want* to use different terminology.
Yes, this sounds nice. Is this possible today already, or will it come
in later versions?
> If 'Active threads' catched more threads, and/or there was an option to
> send a thread there, it would help a lot.
Yes, definately.
> Opera does show a different icon in front of posting from people in
> your contacts, which helps.
It helps, but I try to stay away from the original entry points (e.g.
the newsgroups) as much as possible. Is that stupid of me?
> It would be even better if I could see which threads contain postins
> from people I know without expanding them.
Yes.
> But I think the number of columns is large enough already, and I rather
> keep an uncluttered view of the message list.
But columns could be added and customized, couldn't they? Disabled by
default, but possible to add by right-clicking the column header.
> I find the display in Opera much prettier than TheBat or Agent or
> Eudora or Xnews...
Oh yes, unquestionably.
> But now I'm drifting off-topic :)
Oh yes, unquestionably. ;-)
Asbjørn Ulsberg
> A proper evaluation takes me the better part of a day. That includes a
> pretty extensive regiment of tests to try and "break" the newsreader.
Yes, I thought it looked like a lot of work. Just testing the handling
of the 'References' header took me something like an hour. ;-)
>> Xnews doesn't handle UTF-8 either, but still has the GNKSA seal.
>
> You can blame that on me. I wasn't well versed in UTF-8 when I did the
> Xnews evaluation.
But as UTF-8 support isn't a part of the GNKSA (yet), you didn't do
anything wrong, did you?
>> 'format=flowed' doesn't really mean diddlysquat to me.
>
> It'll mean something to you when you can resize a window and have
> everything still look good.
Neh, not really. I like fixed width in news. I like Courier New as
a news and e-mail font as well. I dislike long lines in these mediums.
>> Maybe we should investigate in how we can get the document updated?
>
> Most GNKSA discussion occurs in news.software.readers.
Ok.
> I've also emailed a pointer to this thread to a couple of the leading
> GNKSA advocates.
Sounds nice.
van Grieg
wrote:
> On Tue, 16 Mar 2004 05:16:04 -0500, van Grieg wrote:
>
>> Unfortunately, while being a wonderful idea overall, Unicode causes a
>> couple of really bad problems. The first one is compatibility.
>
> As Unicode is built to be backward compatible with ASCII, I don't see
> the compatibility-issues. What are they related to, really?
There are programs that can't read and display Unicode.
>> It's still much better to send a few messed up characters than a
>> totally unreadable message.
>
> It's only non-ASCII characters (and not all of those, either) which are
> placed differently in e.g. UTF-8, so as long as the language is written
> in mostly ASCII characters, most of the message will also be readable,
> even without UTF-8 support.
Well, that's the point. Unicode is supposed to benefit those whose
languages _aren't_ written in mostly ASCII characters. But instead it
creates more problems. Just take Russian as an example. Unicode can easily
double the size of the files. And people there are either on dialup or pay
for their broadband traffic. Note also that all programs may handle
encoding detection easily now. So people can read text in their language
without any problems in 99.9% of the cases. Now you introduce Unicode
everywhere, and either they have to wait twice as long for the pages to
load or see their broadband bills grow. Now the question: why in the world
would they like Unicode? For the end user, it doesn't solve any problems
because the problems exist mostly for programmers. Instead they'll just
have to pay more.
Asbjørn Ulsberg
>> As Unicode is built to be backward compatible with ASCII, I don't
>> see the compatibility-issues. What are they related to, really?
>
> There are programs that can't read and display Unicode.
Yes, but in UTF-8 (which is the most common encoding for unicode
characters), all ASCII characters are encoded exactly as in the
ASCII character table, thus will all ASCII characters work with
UTF-8, even if UTF-8 isn't supported.
> Unicode is supposed to benefit those whose languages _aren't_ written
> in mostly ASCII characters.
Yes, but that implies support for Unicode (e.g. UTF-8), of course.
If a client hasn't got support for it, it will display every non-ASCII
character as a glyph, question mark or something similar.
> But instead it creates more problems.
Yes, now it does. But that's just because not all clients support an
Unicode encoding scheme. If they did, it would be much better than the
«every language, its own encoding table!» chaos we have today.
> Just take Russian as an example. Unicode can easily double the size of
> the files.
Yes, it can. But is it really that big of a problem? Aren't excessive
quoting, gigantic signatures and other badly written news articles an
even bigger problem?
> And people there are either on dialup or pay for their broadband traffic.
That's of course something that should go into consideration, but I
think gzip and other solutions to decrease download size are better places
to resort to than to abandon Unicode.
> Note also that all programs may handle encoding detection easily now.
Yes.
> So people can read text in their language without any problems in 99.9%
> of
> the cases. Now you introduce Unicode everywhere, and either they have to
> wait twice as long for the pages to load or see their broadband bills
> grow.
If file size is the only evident problem with Unicode, I say we don't have
a very big problem. If it is the case, then one should initiate work to
get compression to NNTP.
> Now the question: why in the world would they like Unicode? For the end
> user, it doesn't solve any problems because the problems exist mostly for
> programmers. Instead they'll just have to pay more.
If programmers gain, users gain. No question about it. Easier programming
makes easier programs, and even more functional programs. If programmers
don't have to think about encoding problems, users most definately won't.
As it is today, users infact need to know something about encoding if they
are going to exchange data. Maybe most clients will take care of this for
them, but as soon as they run into some encoding problems, either in
received or sent e-mail, they will need to understand what the 'iso-8859-x'
means.
If all applications just used UTF-8 and nothing else, encoding wouldn't
be an issue, neither for the programmer nor for the user. It's the same as
with dates -- earlier you had like 15 different date formats, now more and
more systems are standardizing upon ISO-8601.
Standardizing on one standard (which actually fits as good as all
purposes),
whether it is for character encoding or dates, will make the job easier for
the programmer, which somehow _will_ benefit the user.
Charles Lindsey
>On Mon, 15 Mar 2004 02:16:12 +0100, Rijk van Geijtenbeek wrote:
>> The GNKSA pages [1] seem to be abondened, or at least not updated,
>> for three years now BTW.
Certainly I got no reply from its maintainer the last time I tried to
contact him.
>NNTP hasn't evolved in a lot more years than that. :-) The GNKSA are
>pretty static rules, made to fit the medium NNTP as is. As long as
>the NNTP standard doesn't evolve, there's no apparent reason why the
>GNKSA should either, no?
There is a new NNTP standard in the offing. See
draft-ietf-nntpext-base-21.txt.
>> 2c Provides clear, separate command for replying by e-mail ([2], MUST)
>Yes, this is critical. It's terrible to give users the idea that NNTP
>is the same protocol as POP3 and IMAP, though the messages transfered
>over the protocols is pretty much the same.
Indeed. Reusing Reply and Reply-All as Followup and Post-and-Mail is just
plain sloppiness. Likewise using Send when you mean Post. (Though might it
not be possible to use "Customize toolbars" to fix this?).
>> 3b Warns about, or prevents, posting to large numbers of groups (SHOULD)
>> 13b Allows superseding articles (SHOULD)
>I never do this myself, so not an important issue to me, but a SHOULD is
>a SHOULD is a SHOULD. :-)
What you actually need is a facility to add any arbitrary header to any
article/message. Then people can create the more unusual headers for
themselves (X-No-Archive anyone?).
>> 14a Articles are posted as edited, with linebreaking intact ([4] SHOULD)
>I feel Opera is breaking this a bit, yes.
>> 14b Warns about lines over 80 characters ([4] SHOULD)
>This is a really irritating one. Xnews has a hard break on 80 (or what
>you define) characters, and if you type more than that, the text is
>wrapped to a new line. Also, Xnews has an ugly-looking ruler at the
>top of the edit-area to show where the break is going to occur.
This is where GNKSA needs to catch up with recent developments, like
format=flowed. And 80 is not necessarily the right number in all
circumstances (actually, the proper number should be 79 is you want a hard
number). And the Xnews hard break is certainly wrong (for URLs if nothing
else).
>> 16f Prevents multiple posting entirely ([6] SHOULD)
>Ah, wouldn't this be great.
>> 18a Allows filtering out annoying articles (killing) ([7] SHOULD)
>You can kind of do this with filters, but it's not the same. Both
>scoring and filtering is something I miss deeply in M2 (which I had
>in Xnews).
Yes, that is the biggest single failing in M2. And filters do not really
do the job, because you would need to define as many views as you have
newsgroups.
>> 7d Keeps as many References from original as fit
>I've tested it, and Opera does this quite well. It makes sure that
>the References-header doesn't exceed 998 characters, and starts
>cutting it down when the header reaches the 998-mark.
Better to fold rather than truncate. And Usefor thinking is to trim when
it gets to 21 entries (as opposed to GNKSA which says 3).
>> [4] Opera always uses format=flowed, which obscures the issue being
>> handled with standard 14 "Try to respect the 80-character line-length
>> conventions". By default (but user controllable) a 76 character
>> linelength is enforced, but the compose window does not show this.
>Yup. Not good at all. I hate it, infact.
What it wants is a means to turn format-flowed right off for special
purposes. What we have is just the ability to turn off or alter the
physical line length - useful as far as it goes, but not suficient.
--
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl
Email: c...@clerew.man.ac.uk Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5
Charles Lindsey
>In article <opr4vpta...@news.opera.com>,
> "Rijk van Geijtenbeek" <ri...@opera.com.invalid> wrote:
>> The GNKSA pages [1] seem to be abondened, or at least not
>> updated, for three years now BTW.
>The GNKSA document has not been updated (the USEFOR IETF group[1] is
>currently working on new Usenet standards, to which the GNKSA will
>have to adapt), but there are new evaluations from time to time. There
>aren't that many new newsreaders being developed, and the existing
>ones either aren't updated very often, or already pass the GNKSA.
One of the things the Usefor group is working on is a "Best Practices"
document. One of the issues currently under debate is the extent to which
that should follow or not follow GNKSA.
>[1] http://www.landfield.com/usefor/
--
Richard Grevers
wrote:
>> 14a Articles are posted as edited, with linebreaking intact ([4] SHOULD)
>
> I feel Opera is breaking this a bit, yes.
Or rather, GNKSA has failed to catch up with new standards which have
evolved to address the serious flaws caused by the old standards. Intact
linebreaking, when combined with rigid maximum line lengths, caused ugly,
unreadable "pitchfork quoting", which Format=flowed resolved.
>
>> 14b Warns about lines over 80 characters ([4] SHOULD)
>
> This is a really irritating one. Xnews has a hard break on 80 (or what
> you define) characters, and if you type more than that, the text is
> wrapped to a new line. Also, Xnews has an ugly-looking ruler at the
> top of the edit-area to show where the break is going to occur.
See above.
>
>> 16f Prevents multiple posting entirely ([6] SHOULD)
>
> Ah, wouldn't this be great.
Very difficult to implement. You could search the sent folder and outbox
to see if there is an identical message body, but what if there is one
extra line break or space? How fuzzy do you let matching be? Too far one
way and you fail to prevent multiple posting, too far the other and you
prevent the sending of legitimately different posts.
--
van Grieg
<asbj...@hotmail.com> wrote:
>> Just take Russian as an example. Unicode can easily double the size of
>> the files.
>
> Yes, it can. But is it really that big of a problem?
Yes. See below.
>> And people there are either on dialup or pay for their broadband
>> traffic.
>
> That's of course something that should go into consideration, but I
> think gzip and other solutions to decrease download size are better
> places
> to resort to than to abandon Unicode.
Well, gzip and no Unicode is still better than gzip and Unicode. Gzip is
widely used regardless of charset.
>> Note also that all programs may handle encoding detection easily now.
>
> Yes.
>
>> So people can read text in their language without any problems in 99.9%
>> of
>> the cases. Now you introduce Unicode everywhere, and either they have to
>> wait twice as long for the pages to load or see their broadband bills
>> grow.
>
> If file size is the only evident problem with Unicode, I say we don't
> have
> a very big problem. If it is the case, then one should initiate work to
> get compression to NNTP.
Maybe _you_ don't have a big problem. But what would you say if because of
Unicode you had to pay like $100 more each month? Would you still see it
as a little problem? And remember, it doesn't solve anything for the end
user in most cases, users can read their national languages without
Unicode.
> If all applications just used UTF-8 and nothing else, encoding wouldn't
> be an issue, neither for the programmer nor for the user. It's the same
> as
> with dates -- earlier you had like 15 different date formats, now more
> and
> more systems are standardizing upon ISO-8601.
I understand what you are talking about - I'm just saying there is a
serious downside as well. And once again, encoding is _not_ an issue
really. It's Unicode that introduces new issues when clients don't support
it (I'm talking about non-ASCII languages primarily - for ASCII ones I
don't even see the need for Unicode at all, really). Standards are a great
thing, and Unicode is a great idea. But there's more than just reluctance
to support it - there are solid reasons not to.
Asbjørn Ulsberg
> Or rather, GNKSA has failed to catch up with new standards which have
> evolved to address the serious flaws caused by the old standards. Intact
> linebreaking, when combined with rigid maximum line lengths, caused
> ugly, unreadable "pitchfork quoting", which Format=flowed resolved.
Bad clients do bad formatting. I have _never_ sent a bad formatted news
article from Xnews, because it wraps the text where it will insert breaks
when sending, and you have full control over how your article looks.
Maybe 'format=flowed' solves something, but the real problem lies within
rotten clients like Outlook Express (which needs plugins like OE-QuoteFix
to work properly in this area), not within the standard. Imho, of course.
> Very difficult to implement. You could search the sent folder and outbox
> to see if there is an identical message body, but what if there is one
> extra line break or space?
No, what you do of course, is saving all the generated Message-ID's, so
you can check on them when downloading and indexing new messages.
Frode Gill
wrote:
> On Mon, 15 Mar 2004 02:16:12 +0100, Rijk van Geijtenbeek wrote:
>
>> I know our usenet expert Frode is well aware of GNSKA. He pointed out
>> Opera's failures before he joined the company :)
>
> Goodie. I know he used Xnews earlier and are aware of stuff like this,
> but I feel that he has sunken into some kind of void inside Opera,
> and that he doesn't get the time he needs to make Opera the best news-
> reader ever. I'm sure that's what he'd do, if he had the time.
Your comments are very accurate. I was hired to do NNTP, but after having
worked on POP and SMTP, I am now re-implementing IMAP, and will have to go
through PGP before I can get to what I want - improve the NNTP code.
Please be patient :-/
--
Frode Gill
Rijk van Geijtenbeek
..
> I understand what you are talking about - I'm just saying there is a
> serious downside as well. And once again, encoding is _not_ an issue
> really. It's Unicode that introduces new issues when clients don't
> support it (I'm talking about non-ASCII languages primarily - for ASCII
> ones I don't even see the need for Unicode at all, really).
Using UTF-8 makes it possible to unclude a few Russian and Greek words in
a text. For multi-language texts, it is very useful. From [1]:
"Have you ever tried to include a passage in a different alphabet in one
of your documents, for example a quotation in Russian in an English
document, only to find that you have no Cyrillic characters available? Or
sent a Spanish document in electronic form to someone in Greece, only to
be told that the accented Latin characters have been replaced by Greek
characters? Or produced a Web page that includes technical symbols and
found that it works with Windows but not with Mac OS or Unix? Problems
like these arise with non-Latin alphabets and Symbol fonts because until
recently most computers used fonts that contain a maximum of 256
characters."
I see your point about growing file sizes of course. But newsreaders that
are based on using the Unicode character set internally, shoudl have no
problem at all in receiving and sending all kinds of encodings, provided
they are labeled correctly, and the user can choose the best one for his
cause. And with manual overrides in case the labeling was wrong, you can
read everything :)
Non-unicode aawre newsreaders will have more trouble. I remember the
trouble I had when I wanted to include the euro symbol in a message posted
with Agent. And I wouldn't know how to send the smiley like I did in this
thread, as an example of characters that fall outside the normal encodings.
[1] http://www.alanwood.net/unicode/index.html
Rijk van Geijtenbeek
wrote:
..
> OMG! It trims leading space in every other scenario (damaging the
> signature separator '-- \n'), but here it adds it. Great! :-\
I just started reading news.software.readers, and found that recent
versions of OE don't damage the sig separator anymore :)
Rijk van Geijtenbeek
wrote:
[..]
>>> 2c Provides clear, separate command for replying by e-mail ([2], MUST)
>
>> Yes, this is critical. It's terrible to give users the idea that NNTP
>> is the same protocol as POP3 and IMAP, though the messages transfered
>> over the protocols is pretty much the same.
>
> Indeed. Reusing Reply and Reply-All as Followup and Post-and-Mail is just
> plain sloppiness. Likewise using Send when you mean Post. (Though might
> it not be possible to use "Customize toolbars" to fix this?).
As I said elsewhere, there is no separate toolbar for newsgroups, it
shares with the other views. Just as the Compose window is the same
(though with an appropriate selection of headers fields to fill in) as
that for mail messages.
It is *not* a goal of Opera to make the newsreader UI as much as possible
the same as other newsreaders. Instead, offering seamless integration for
all sorts of messages is the goal. Hence the inclusion even of newsfeeds
in the same panel in 7.5, with the same toolbar.
People coming from other newsreaders might frown when seeing this toolbar,
but people used to mail and trying out newsgroups for the first time will
be very happy OTOH. I don't think Opera breaks the spirit of the GNKSA
here, as the buttons available do not unexpectedly cause disruptive
behavior.
I know you will disagree :)
[..]
Rijk van Geijtenbeek
wrote:
> On Wed, 17 Mar 2004 17:11:35 +1300, Richard Grevers wrote:
I think you mix two issues here: catching incoming duplicates and
preventing multi-posting. I don't think Opera needs to do better in the
latter, but there is certainly room for improvement on the former. But
that falls outside GNKSA [16 e].
IMHO, multi-posting should not be completely impossible. For example, the
news.opera.com rejects articles crossposted to groups on another server.
But I might went to to send exciting news on Opera's CSS support to
c.i.w.a.s as well as opera.page-authoring (and to the CSS-Discuss mailing
list). Why should I be pestered with a popup warning when I open the
message again from the Sent view? Is this likely to be done casually by
clueless newbies? GNKSA is about helping those new users to conform to
netiquette, after all.
Asbjørn Ulsberg
> I just started reading news.software.readers, and found that recent
> versions of OE don't damage the sig separator anymore :)
True. Microsoft released a patch for this in April 2003, but still
most Outlook Express'es aren't patched properly. OE still, after
the patch, encourages top quoting afaik.
Asbjørn Ulsberg
> Your comments are very accurate. I was hired to do NNTP, but after
> having worked on POP and SMTP, I am now re-implementing IMAP, and will
> have to go through PGP before I can get to what I want - improve the
> NNTP code. Please be patient :-/
I'll be patient. But can you try to give me a stick to hold on to?
Something concrete? Something to look forward to (like a date)? :-)
Asbjørn Ulsberg
> I think you mix two issues here: catching incoming duplicates and
> preventing multi-posting.
Yes, I probably am.
> I don't think Opera needs to do better in the latter
I dunno how Opera behaves on this right now, but it's probably not
terrible.
> but there is certainly room for improvement on the former. But that
> falls outside GNKSA [16 e].
Yes it does, but can still be improved, no?
> IMHO, multi-posting should not be completely impossible.
No, but a warning message should imho be provided (which you could
turn off permanently with «Do not ask me this again»).
Charles Lindsey
>On Tue, 16 Mar 2004 02:07:15 +0100, Asbjørn Ulsberg <asbj...@hotmail.com>
>wrote:
>> I don't fully agree with that. Even with 'format=flowed', I try to make
>> my articles pretty, keep the lines within 80 characters, etc. I think
>> this is something good, at least that I _can_ control this. I don't think
>> I have enough control in Opera, and it makes me very anxious.
>Opera keeps the lines within 76 characters automatically, and shows what
>it does with quoted text.
72 characters would be better (or 74 at a pinch) so as to allow room for
quoting by others who do not have the marvels of f=f on their systems. I
have configured in thus in index.ini.
>> My problem lies in the fact that news and e-mail isn't clearly enough
>> separated in M2. That's what I'm talking about.
>For Opera, it is a feature that mail and news are tightly integrated.
>Don't expect chganges on that front. It should be rather easy to have a
>separate toolbar for newsgroup windows, which would allow the use of
>special terminology ('followup', 'reply-by-mail'). But I don't think we
>really *want* to use different terminology.
Yes you do. That is precisely why GNKSA made an issue of it, because it
was observed that people who tried to graft newsreading on to the top of
mail readers (which usually leads to a pretty poor newsreader) were taking
too many shortcute like that. In fact, you need 3 buttins for news:
"followup", "reply" and "followup + reply" (though the latter might be
provided by some extra button in the compose window on the other two -
indeed, GNKSA asks for a feature to change between those three even when
you are in the midst of composing).
Charles Lindsey
>On Tue, 16 Mar 2004 09:48:36 +0100, Rijk van Geijtenbeek wrote:
>> Opera keeps the lines within 76 characters automatically, and shows what
>> it does with quoted text.
>I don't think it shows clearly enough. I want a ruler, goddamnit! #:-D
What I have done is to adjust the size of my mail and compose windows to
be around 80 characters. That gives a reasonable indication of how it is
going to look in the usual 79 character viewing window.
Oh, and I have to fill my signature (see below) with non-breaking spaces
just to stop Opera from trying to fold it :-(.
Rijk van Geijtenbeek
wrote:
> In <opr4x5fa...@news.opera.com> "Rijk van Geijtenbeek" writes:
>> For Opera, it is a feature that mail and news are tightly integrated.
>> Don't expect chganges on that front. It should be rather easy to have a
>> separate toolbar for newsgroup windows, which would allow the use of
>> special terminology ('followup', 'reply-by-mail'). But I don't think we
>> really *want* to use different terminology.
>
> Yes you do. That is precisely why GNKSA made an issue of it, because it
> was observed that people who tried to graft newsreading on to the top of
> mail readers (which usually leads to a pretty poor newsreader) were
> taking too many shortcute like that.
Those shortcuts mean we now have a nice little newsreader, very usable for
text based newsgroups, that doesn't take up much resources. And I still
don't understand why 'followup' is somehow easier to understand than
'reply'. For new users, it might even get confusing when people tell them
"don't forget to set a followup" after to do their first crossposting...
> In fact, you need 3 buttins for news:
> "followup", "reply" and "followup + reply" (though the latter might be
> provided by some extra button in the compose window on the other two -
> indeed, GNKSA asks for a feature to change between those three even when
> you are in the midst of composing).
Opera already makes that very easy: when you press 'Reply all', you have
'To', 'News' and 'Followup' fields at your disposal by default. So it is
very easy to delete either the 'To' or the 'News' entry that is filled in.
If you start by pressing 'Reply', you don't get the 'To' field, but you
can add it if you want. Would you want to get a button to add a 'To' field
and fill in the from/reply-to address for you?
Richard Grevers
wrote:
>
> People coming from other newsreaders might frown when seeing this
> toolbar, but people used to mail and trying out newsgroups for the first
> time will be very happy OTOH. I don't think Opera breaks the spirit of
> the GNKSA here, as the buttons available do not unexpectedly cause
> disruptive behavior.
>
All it needs is the adding of two items to the skin files and
customisation dialogue: followup (usung the same image as reply), and a
new image for (email) reply to author. Even if these are not defaults, the
ability to set them would be nice.
--
Richard Grevers
wrote:
> On Wed, 17 Mar 2004 19:19:49 GMT, Charles Lindsey <c...@clerew.man.ac.uk>
> wrote:
>> In <opr4x5fa...@news.opera.com> "Rijk van Geijtenbeek" writes:
>
>>> For Opera, it is a feature that mail and news are tightly integrated.
>>> Don't expect chganges on that front. It should be rather easy to have a
>>> separate toolbar for newsgroup windows, which would allow the use of
>>> special terminology ('followup', 'reply-by-mail'). But I don't think we
>>> really *want* to use different terminology.
>>
>> Yes you do. That is precisely why GNKSA made an issue of it, because it
>> was observed that people who tried to graft newsreading on to the top of
>> mail readers (which usually leads to a pretty poor newsreader) were
>> taking too many shortcute like that.
>
> Those shortcuts mean we now have a nice little newsreader, very usable
> for text based newsgroups, that doesn't take up much resources. And I
> still don't understand why 'followup' is somehow easier to understand
> than 'reply'. For new users, it might even get confusing when people
> tell them "don't forget to set a followup" after to do their first
> crossposting...
>
Well we seem to have plenty of confused opera-newbies. Perhaps like me
they came from several years of using standard news clients where "reply"
means "reply to author by email" (and is usually the button to the right
of "followup".
Asbjørn Ulsberg
>> I don't think it shows clearly enough. I want a ruler, goddamnit! #:-D
>
> What I have done is to adjust the size of my mail and compose windows to
> be around 80 characters.
That could work, and was the half-ass way I solved it in Outlook Express.
But I don't want to use OE hacks on Opera! That's just sad, man! :-\
> That gives a reasonable indication of how it is going to look in the
> usual
> 79 character viewing window.
Yes, but it's terrible. I like maximized windows. Don't ask me why, I just
do. And Opera shouldn't force me away from that fetishism. :-)
> Oh, and I have to fill my signature (see below) with non-breaking spaces
> just to stop Opera from trying to fold it :-(.
I'm not sure what you mean, here. Care to elaborate?
Asbjørn Ulsberg
> 72 characters would be better (or 74 at a pinch) so as to allow room for
> quoting by others who do not have the marvels of f=f on their systems. I
> have configured in thus in index.ini.
How?
>> [...] allow the use of special terminology ('followup',
>> 'reply-by-mail').
>> But I don't think we really *want* to use different terminology.
>
> Yes you do.
I agree!
> That is precisely why GNKSA made an issue of it, because it was observed
> that people who tried to graft newsreading on to the top of mail readers
> (which usually leads to a pretty poor newsreader) were taking too many
> shortcute like that.
Yup. It's a different action on a different medium. Opera should reflect,
not conceal that fact.
> In fact, you need 3 buttins for news: "followup", "reply" and "followup +
> reply"
The latter should imho be hidden by default, but could be possible to drag/
drop into the toolbar with «Customize toolbars».
Asbjørn Ulsberg
> Those shortcuts mean we now have a nice little newsreader, very
> usable for text based newsgroups, that doesn't take up much
> resources.
You think adding a couple of more buttons to M2 will explode its size
and resource usage out of proportions? ;-)
> And I still don't understand why 'followup' is somehow easier to
> understand than 'reply'.
It's maybe not easier to understand 'followup', but 'reply' already
has a defined meaning for most people through their e-mail clients.
Hence, introducing a new meaning (on a different medium) isn't wise.
'reply' and 'followup' do different things, and 'reply' should, if
available in the news part of a client, do the exact same thing as
in the e-mail part. It should start on a new message which will be
sent to the author of the message, by e-mail.
If you dislike the word 'followup', I can totally live with two
buttons saying 'Reply to group' and 'Reply to author' or something
similar. I don't think GNKSA's mission here is to force everyone
to use NNTP's technical jargon, but to have different buttons
doing different things for different actions, not «one button that
does it all». :-p
> Opera already makes that very easy: when you press 'Reply all',
> you have 'To', 'News' and 'Followup' fields at your disposal by
> default.
I think this is horrific. Sorry, but I do. 'Reply all' is a nice
function, but is _not_ the button I would press if I wanted to
reply to the news article's author.
> So it is very easy to delete either the 'To' or the 'News' entry
> that is filled in.
I can't help thinking «This is a hack. This is a hack. This is a
hack». ;-)
> If you start by pressing 'Reply', you don't get the 'To' field, but you
> can add it if you want. Would you want to get a button to add
> a 'To' field and fill in the from/reply-to address for you?
No, I think what we want are two separate buttons for 'reply to
author' and 'followup (or whatever) to newsgroup'.
Asbjørn Ulsberg
> Well, gzip and no Unicode is still better than gzip and Unicode.
gzip and no Unicode makes smaller messages than gzip and Unicode,
but I don't agree that it's better. See Rijk's reply on this.
> Gzip is widely used regardless of charset.
On NNTP? Where? And how?
> Maybe _you_ don't have a big problem. But what would you say if
> because of Unicode you had to pay like $100 more each month?
I would scream so loud and hard that my vocal chords would be
flung out of my throat and stuck on the wall. But I really don't
believe this. Sorry.
> Would you still see it as a little problem?
No, but I don't believe the scenario. I don't think NNTP can ever
cause that much traffic on anyone's line, and if someone only uses
their connection to read and post Usenet articles, they have
bigger problems than the size of the bytes on the wire! :-p
> And remember, it doesn't solve anything for the end user in most
> cases, users can read their national languages without Unicode.
They can read their national languages, yes, but often not other's.
And never combined languages, as Rijk points out.
> I understand what you are talking about - I'm just saying there is
> a serious downside as well.
There is a downside, yes, but I don't think it's that serious. I'm
sorry to say I believe you're exaggerating a bit.
> And once again, encoding is _not_ an issue really. It's Unicode
> that introduces new issues when clients don't support it
Yes, but clients _should_ support it. In 2004, there's virtually
_no_ reason for _any_ software not to support Unicode. Really.
> Standards are a great thing, and Unicode is a great idea. But
> there's more than just reluctance to support it - there are solid
> reasons not to.
I digress, but there's always room for different opinions, no? :-)
Rijk van Geijtenbeek
wrote:
> On Wed, 17 Mar 2004 19:19:49 GMT, Charles Lindsey wrote:
>
>> 72 characters would be better (or 74 at a pinch) so as to allow room for
>> quoting by others who do not have the marvels of f=f on their systems. I
>> have configured in thus in index.ini.
>
> How?
I think Charles meant 'accounts.ini'. Close Opera, open 'accounts.ini'
(found in the \Mail subdirectory), search for '76' and replace with '72' :)
..
> Yup. It's a different action on a different medium. Opera should reflect,
> not conceal that fact.
Why?
Rijk van Geijtenbeek
wrote:
> On Thu, 18 Mar 2004 10:10:33 +0100, Rijk van Geijtenbeek wrote:
>
>> Those shortcuts mean we now have a nice little newsreader, very
>> usable for text based newsgroups, that doesn't take up much
>> resources.
>
> You think adding a couple of more buttons to M2 will explode its size
> and resource usage out of proportions? ;-)
The newsreader is a add-on grafted on the M2 client. Apart from the
support for a different network protocol, this is very little coding done
it. This is a pity, but if it couldn't be done this way it might not have
been done at all... So that means little resources compared to opening
both Opera and, say, Agent at the same. As I used to do before Opera 7.
..
van Grieg
wrote:
>> Gzip is widely used regardless of charset.
>
> On NNTP? Where? And how?
Oh, not on NNTP, sure. And yes, GZIP would help.
>> Maybe _you_ don't have a big problem. But what would you say if
>> because of Unicode you had to pay like $100 more each month?
>
> I would scream so loud and hard that my vocal chords would be
> flung out of my throat and stuck on the wall. But I really don't
> believe this. Sorry.
Nothing to be sorry about. $100 is what extra 0.5-2 GB of traffic in
Russia will cost you, depending on location. That's life. This is
Walgreens. And let's not forget about people on dialup.
>> Would you still see it as a little problem?
>
> No, but I don't believe the scenario. I don't think NNTP can ever
> cause that much traffic on anyone's line, and if someone only uses
> their connection to read and post Usenet articles, they have
> bigger problems than the size of the bytes on the wire! :-p
NNTP alone isn't likely to generate so much traffic. But it all adds up at
the end of the day. No Unicode anywhere is better in this sense.
>> And remember, it doesn't solve anything for the end user in most
>> cases, users can read their national languages without Unicode.
>
> They can read their national languages, yes, but often not other's.
> And never combined languages, as Rijk points out.
Sure. The question is: how much is a smilie? And the bad thing about it is
that the recepient can't choose. It's exactly the same problem as with
HTML mail, IMO. It's an embellishment, not a necessity.
>
>> And once again, encoding is _not_ an issue really. It's Unicode
>> that introduces new issues when clients don't support it
>
> Yes, but clients _should_ support it. In 2004, there's virtually
> _no_ reason for _any_ software not to support Unicode. Really.
I agree 100%. There _are_ occasions when it may be useful, even though
they are rare. But going back to requirements for newsreaders, I just
don't see Unicode used anywhere on NNTP, so it's not a big deal anyway.
Just my 2 cents. I absolutely don't mind if programs support it. But if
they don't, it doesn't worry me too much either.
Rijk van Geijtenbeek
wrote:
..
>> And remember, it doesn't solve anything for the end user in most
>> cases, users can read their national languages without Unicode.
>
> They can read their national languages, yes, but often not other's.
> And never combined languages, as Rijk points out.
I just read in n.s.r. about people having to use a proxy to be able to use
XNews with anything but US-Ascii...
Unicode provides a great framework to build truly international
applications. Van Grieg has a good point that this doesn't make UTF-8 the
best encoding for all languages. It is a good encoding for multi-language
texts though.
..
>> Standards are a great thing, and Unicode is a great idea. But
>> there's more than just reluctance to support it - there are solid
>> reasons not to.
There are good reasons to only *send* as UTF-8 when you really need to,
and use more efficient encodings when possible. Especially when you know
that many recipients use older software.
I fail to see any good reason for applications not to support as many
encodings as possible, using Unicode as a framework. But implementing
Unicode support is quite a lot of work when you want to do it for older
operating systems as Windows 9x, as Opera's developers can testify...
van Grieg
Just to clarify my position a bit: I don't oppose the idea of Unicode as
such. I just call on people to be cautious in using it. Just like many
great technologies, I'm afraid it may be easily abused. I like the way
Opera deals with it, by the way. It's not used by default, but can be
triggered if needed (and now there's a warning, so I can turn it off if I
need to). I think _such_ behavior would make a good standard.
van Grieg
wrote:
> There are good reasons to only *send* as UTF-8 when you really need to,
> and use more efficient encodings when possible. Especially when you know
> that many recipients use older software.
Absolutely.
> I fail to see any good reason for applications not to support as many
> encodings as possible, using Unicode as a framework. But implementing
> Unicode support is quite a lot of work when you want to do it for older
> operating systems as Windows 9x, as Opera's developers can testify...
>
And, as may be seen on Opera's example, this may also greatly increase the
size of the application. Not that it matters now for most people, but
still. But overall, yes, I wouldn't consider a mail client or a newsreader
(for me as an Opera user they have long ago become the same thing) without
Unicode support.
van Grieg
> I wouldn't consider a mail client or a newsreader (for me as an Opera
> user they have long ago become the same thing) without Unicode support.
>
I should get more sleep for sure. This is not a contradiction to my
earlier post where I say that if clients didn't support Unicode I wouln't
worry about it. It's just that if there _are_ clients that do, I'd rather
choose among them.
Asbjørn Ulsberg
>> You think adding a couple of more buttons to M2 will explode its size
>> and resource usage out of proportions? ;-)
>
> The newsreader is a add-on grafted on the M2 client. Apart from the
> support for a different network protocol, this is very little coding
> done it.
I've understood that. And Frode says his focus is on IMAP now. Who the
frick cares about IMAP!? :p
> This is a pity, but if it couldn't be done this way it might not have
> been done at all...
To even consider not implementing NNTP makes as much sense to me as
using a running chainsaw as a comb. :-)
> So that means little resources compared to opening both Opera and,
> say, Agent at the same.
True. In this field, Opera rocks. But a method of getting more out of
the newsreader might just be some downloadable component for what I
care. I prioritize functionality over size, startup- and download-
time.
But I expect to see drastic improvements in the news client of M2
when Frode is back in the chair he belongs, hacking on the über
protocol of them all; NNTP. #:-D
Asbjørn Ulsberg
> I just read in n.s.r. about people having to use a proxy to be able
> to use XNews with anything but US-Ascii...
Xnews supports ISO-8859-1, but not much else. E.g., it supports what
Delphi supports out of the box. Which isn't much. I installed a little
proxy called 'mproxy' to filter and translate all Unicode and such
into ISO-8859-1, so that Xnews would understand it.
It worked like a charm, but at some point I had 4 services running
just to please Xnews, and I found that it had to be an easier way. So
I installed Opera 7.5, and are almost happy with it. Almost.
> Unicode provides a great framework to build truly international
> applications.
Definately.
> Van Grieg has a good point that this doesn't make UTF-8 the best
> encoding for all languages. It is a good encoding for multi-language
> texts though.
Yes. Use the smallest possible encoding scheme for what you send, but
support the widest and biggest (i.e. «everything») for what you
receive.
> There are good reasons to only *send* as UTF-8 when you really need to,
> and use more efficient encodings when possible. Especially when you know
> that many recipients use older software.
Yes. It's a de facto standard in the Norwegian newsgroups I frequent to
only send messages in ISO-8859-1. Mostly because no other encoding is
required for the characters we use. But I feel that UTF-8 does no big
harm, at least not regarding bytesize.
> I fail to see any good reason for applications not to support as many
> encodings as possible, using Unicode as a framework.
Me neither.
> But implementing Unicode support is quite a lot of work when you want to
> do it for older operating systems as Windows 9x, as Opera's developers
> can testify...
Oh, I'm so glad I don't have to do stuff like that. Soooo glad!
Asbjørn Ulsberg
>> How?
>
> I think Charles meant 'accounts.ini'. Close Opera, open 'accounts.ini'
> (found in the \Mail subdirectory), search for '76' and replace with '72'
Oh. I thought maybe it was possible to get Opera to wrap the content
edit window, now. Bah! :-)
>> Yup. It's a different action on a different medium. Opera should
>> reflect, not conceal that fact.
>
> Why?
Because the user needs to know the difference. If a tool gives you the
impression that HTML is a visual markup langauge, you use it as such.
But if it helps you understand that it is a structural markup language,
you might have more success with using it the way it's meant to be
used.
What I'm trying to say with this stupid analogy is that people don't
benefit anything from being deceived by their software. If the software
is frank, honest and tells the user what's what and nothing else, the
user has a better platform for understanding how the given technology
actually works.
So concealing the fact that NNTP != POP3 doesn't benefit the user.
But implementing both protocols in an as seamless way as possible,
does. There's a difference, and currently Opera does both.
J.B. Moreno
> On Mon, 15 Mar 2004 20:03:43 -0600, Adam Bailey wrote:
> >> 'format=flowed' doesn't really mean diddlysquat to me.
> >
> > It'll mean something to you when you can resize a window and have
> > everything still look good.
>
> Neh, not really. I like fixed width in news. I like Courier New as
> a news and e-mail font as well. I dislike long lines in these mediums.
Even with a fixed width, format=flowed has it's advantages -- you don't
have to keep manually rewrapping the quoted text, and when snipping,
things just end up looking good.
Not that you can't approximate this without format=flowed, but it's
more error prone.
Also, there is nothing in format=flowed that says you can't wrap the
lines at a fixed width for display -- I think that'd be a good idea as
a default as I recently told someone else implementing f=f.
--
J.B. Moreno
J.B. Moreno
> Rijk van Geijtenbeek wrote:
> > So it is very easy to delete either the 'To' or the 'News' entry
> > that is filled in.
>
> I can't help thinking «This is a hack. This is a hack. This is a
> hack». ;-)
Yeah. I normally use a Mac, and in MacSOUP, it's a nice little drop
down menu in the upper left of the compose window
Email Reply (private)
Followup (public)
Followup & Email Copy
I also use Thoth, where the first 2 "push" buttons in the top left,
"News" (text description when you hover "Post the message via usenet")
and a mail envelope (text description "Send the message via email").
You can switch back and forth between the 3 states in either of these
program in under a second.
--
J.B. Moreno
J.B. Moreno
> Certainly I got no reply from its maintainer the last time I tried to
> contact him.
I just submitted a preliminary evaluation of Mozilla, and got a prompt
response.
--
J.B. Moreno
Asbjørn Ulsberg
> Even with a fixed width, format=flowed has it's advantages -- you don't
> have to keep manually rewrapping the quoted text, and when snipping,
> things just end up looking good.
I don't have to rewrap manually in Xnews either. I want my newsreader to
rewrap automatically, but I want it to do it as soon as frikkin' possible.
It seems as Opera does a combination -- it rewraps when replying (after
adding the quoting character '>'), and then again when I send.
This is OK, but when I edit the text, I get paranoid from not seeing
exactly how it's going to be, but that might just be somethin I have to
get used to.
> Not that you can't approximate this without format=flowed, but it's
> more error prone.
Yes, it probably is.
> Also, there is nothing in format=flowed that says you can't wrap the
> lines at a fixed width for display -- I think that'd be a good idea as
> a default as I recently told someone else implementing f=f.
Yes, and this is exactly what I want from Opera. I'm sure it's possible
to do almost entirely with CSS (max-width: 72em), so give me a ruler,
and show me the wrapping! Now! :-p
J.B. Moreno
> On Thu, 18 Mar 2004 21:29:06 -0500, J.B. Moreno wrote:
>
> > Even with a fixed width, format=flowed has it's advantages -- you don't
> > have to keep manually rewrapping the quoted text, and when snipping,
> > things just end up looking good.
>
> I don't have to rewrap manually in Xnews either.
You do when you edit the quoted text -- i.e. when snipping or inserting
a comment as I've done here.
> I want my newsreader to rewrap automatically, but I want it to do it
> as soon as frikkin' possible. It seems as Opera does a combination --
> it rewraps when replying (after adding the quoting character '>'),
> and then again when I send.
>
> This is OK, but when I edit the text, I get paranoid from not seeing
> exactly how it's going to be, but that might just be somethin I have to
> get used to.
Well, I obviously can't say whether you'll have to get used to it or
not, but IMO it's certainly a reasonable wish -- I like knowing how my
post looks as well.
> > Not that you can't approximate this without format=flowed, but it's
> > more error prone.
>
> Yes, it probably is.
I wrote the rewrapping code for Xnews (and then Luu greatly extended
it), so take my word for it -- it's more error prone.
> > Also, there is nothing in format=flowed that says you can't wrap the
> > lines at a fixed width for display -- I think that'd be a good idea as
> > a default as I recently told someone else implementing f=f.
>
> Yes, and this is exactly what I want from Opera. I'm sure it's possible
> to do almost entirely with CSS (max-width: 72em), so give me a ruler,
> and show me the wrapping! Now! :-p
I'm not sure if Mozilla does this via CSS or not, but it's plain text
editor does it for unquoted text (quoted text can be longer).
--
J.B. Moreno
Rijk van Geijtenbeek
wrote:
> Asbjørn Ulsberg <asbj...@hotmail.com> wrote:
>
>> On Thu, 18 Mar 2004 21:29:06 -0500, J.B. Moreno wrote:
>>
>> > Even with a fixed width, format=flowed has it's advantages -- you
>> don't
>> > have to keep manually rewrapping the quoted text, and when snipping,
>> > things just end up looking good.
Things went wrong with the quoting in the piece above... If you'd all use
f=f software, it would look like this:
>>> Even with a fixed width, format=flowed has it's advantages -- you
>>> don't have to keep manually rewrapping the quoted text, and when
>>> snippingthings just end up looking good.
... without any manual action. If the line gets to long, you get the
pitchfork effect, but with the right number of > characters in front.
Which displays fine in other f=f readers.
Thoth is breaking some spec here by adding a space between the new > and
the original >, which causes the quote levels to break.
>> I don't have to rewrap manually in Xnews either.
>
> You do when you edit the quoted text -- i.e. when snipping or inserting
> a comment as I've done here.
>
>> I want my newsreader to rewrap automatically, but I want it to do it
>> as soon as frikkin' possible. It seems as Opera does a combination --
>> it rewraps when replying (after adding the quoting character '>'),
>> and then again when I send.
It wraps the quotes directly (as displayed above, it does better when
proper > characters are used). If you don't add anything to the quoted
lines, no change will occur there.
>> This is OK, but when I edit the text, I get paranoid from not seeing
>> exactly how it's going to be, but that might just be somethin I have to
>> get used to.
>
> Well, I obviously can't say whether you'll have to get used to it or
> not, but IMO it's certainly a reasonable wish -- I like knowing how my
> post looks as well.
The breaking of your typed reply at 76 characters? Why do you care, unless
composing ASCII art?
If I try to manually improve the pitchfork quotes like this:
>> Very long sentence very long sentence very long sentence
>> very long sentence very long sentence very long sentence
>> very
>> long sentence very long sentence very long sentence very
>> long sentence ends here.
.. it would indeed be nice if I could be sure of what I did, and a visible
helpline would be nice then.
[..]
>>> Also, there is nothing in format=flowed that says you can't wrap the
>>> lines at a fixed width for display -- I think that'd be a good idea as
>>> a default as I recently told someone else implementing f=f.
>>
>> Yes, and this is exactly what I want from Opera. I'm sure it's possible
>> to do almost entirely with CSS (max-width: 72em), so give me a ruler,
>> and show me the wrapping! Now! :-p
>
> I'm not sure if Mozilla does this via CSS or not, but it's plain text
> editor does it for unquoted text (quoted text can be longer).
CSS doesn't have a length value that counts characters, alas. The values
'em' or 'ex' are not really suited for this. You could use trial and error
to see which 'em' value would work for the current monospaced font you
use, but it would be different for each font.
Try to add this to your mime.css:
omf|body, omf|div {max-width: 35em}
I've posted a ruler to opera.test :)
Brian D Clary
wrote:
> On Thu, 18 Mar 2004 22:11:36 +0100, Rijk van Geijtenbeek wrote:
>
>>> You think adding a couple of more buttons to M2 will explode its size
>>> and resource usage out of proportions? ;-)
>>
>> The newsreader is a add-on grafted on the M2 client. Apart from the
>> support for a different network protocol, this is very little coding
>> done it.
>
> I've understood that. And Frode says his focus is on IMAP now. Who the
> frick cares about IMAP!? :p
Me :p (If the poor guy has to rewrite the IMAP again, then I think he
should know that at least someone appreciates it.)
>> This is a pity, but if it couldn't be done this way it might not have
>> been done at all...
>
> To even consider not implementing NNTP makes as much sense to me as
> using a running chainsaw as a comb. :-)
>
>> So that means little resources compared to opening both Opera and,
>> say, Agent at the same.
>
> True. In this field, Opera rocks. But a method of getting more out of
> the newsreader might just be some downloadable component for what I
> care. I prioritize functionality over size, startup- and download-
> time.
For me, it's whatever gets the job done easiest. And despite my usual
leanings that specialization works better is most cases, it's easier for
me now to use Opera for email, news, and browsing, than to open Pine
alongside Opera. And that was hard to do, as I'm a big Pine fan; I still
use on it occasion when I want to see exactly what's in my IMAP folders
with their IMAP flags. Though I should mention that M2's
filters/views/whatevers had a big part to do with it also.
> But I expect to see drastic improvements in the news client of M2
> when Frode is back in the chair he belongs, hacking on the über
> protocol of them all; NNTP. #:-D
To each his own ;)
--
Brian
Opera 7.50 build 3658
Windows XP Pro
To email, replace domain with operamail(.com)
J.B. Moreno
> J.B. Moreno <pl...@newsreaders.com> wrote:
>
> > Asbj¯rn Ulsberg <asbj...@hotmail.com> wrote:
> >
> >> On Thu, 18 Mar 2004 21:29:06 -0500, J.B. Moreno wrote:
> >>
> >> > Even with a fixed width, format=flowed has it's advantages -- you
> >> don't
> >> > have to keep manually rewrapping the quoted text, and when snipping,
> >> > things just end up looking good.
>
> Things went wrong with the quoting in the piece above... If you'd all use
> f=f software, it would look like this:
It went wrong on your end -- it was good as it went out (which I just
verified).
> >>> Even with a fixed width, format=flowed has it's advantages -- you
> >>> don't have to keep manually rewrapping the quoted text, and when
> >>> snippingthings just end up looking good.
>
> ... without any manual action. If the line gets to long, you get the
> pitchfork effect, but with the right number of > characters in front.
> Which displays fine in other f=f readers.
>
> Thoth is breaking some spec here by adding a space between the new > and
> the original >, which causes the quote levels to break.
No, it's Opera messing up -- Thoth neither added nor removed any
spaces, Opera did: "> > Even" is how I sent it, "> >> > Even" is how
it was in your followup. Notice that it both added an extra level of
indentation *and* removed a space.
The later (removing the space) is not uncommon, particularly with
programs that support format=flowed (by changing all instances of "> "
to just ">", they can have the text flow using their regular routines).
-snip-
> >> This is OK, but when I edit the text, I get paranoid from not seeing
> >> exactly how it's going to be, but that might just be somethin I have to
> >> get used to.
> >
> > Well, I obviously can't say whether you'll have to get used to it or
> > not, but IMO it's certainly a reasonable wish -- I like knowing how my
> > post looks as well.
>
> The breaking of your typed reply at 76 characters? Why do you care, unless
> composing ASCII art?
Two reasons -- (1) paragraph presentation isn't just for ASCII art, and
more even lines are less jarring, and (2) bad experience with programs
doing the wrong thing, I like to know what's going out, then if the
program is messing up, I have a chance of catching it myself.
-snip-
> >>> Also, there is nothing in format=flowed that says you can't wrap the
> >>> lines at a fixed width for display -- I think that'd be a good idea as
> >>> a default as I recently told someone else implementing f=f.
> >>
> >> Yes, and this is exactly what I want from Opera. I'm sure it's possible
> >> to do almost entirely with CSS (max-width: 72em), so give me a ruler,
> >> and show me the wrapping! Now! :-p
> >
> > I'm not sure if Mozilla does this via CSS or not, but it's plain text
> > editor does it for unquoted text (quoted text can be longer).
>
> CSS doesn't have a length value that counts characters, alas. The values
> 'em' or 'ex' are not really suited for this. You could use trial and error
> to see which 'em' value would work for the current monospaced font you
> use, but it would be different for each font.
If CSS doesn't do it, then you have to do it some other way -- either
that or extend CSS, maybe that's what Mozilla did.
--
J.B. Moreno
Charles Lindsey
>On Wed, 17 Mar 2004 19:23:22 GMT, Charles Lindsey wrote:
>> Oh, and I have to fill my signature (see below) with non-breaking spaces
>> just to stop Opera from trying to fold it :-(.
>I'm not sure what you mean, here. Care to elaborate?
The lines in my signature are exactly 79 characters long (see below).
When you bring up the Compose window, Opera preloads my signature at the
bottom. Fine!
But when it sends the message, it observes that the length of the lines is
longer than 76 characters (or 74 characters in my setup). So it splits
each line in two before sending it. Those who read it in flowed-enabled
readers will see it restored to its former glory. Those who read it in
conventional readers will see it chopped up.
So, to prevent that happening, I have had to replace all the space
characters by non-breaking spaces.
Note that you will not see that in the signature below, because I am not
posting this through Opera.
--
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl
Email: c...@clerew.man.ac.uk Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5
Charles Lindsey
>On Wed, 17 Mar 2004 19:19:49 GMT, Charles Lindsey wrote:
>> In fact, you need 3 buttins for news: "followup", "reply" and "followup +
>> reply"
>The latter should imho be hidden by default, but could be possible to drag/
>drop into the toolbar with «Customize toolbars».
And if you were to use "followup + reply", it ought to look whether there
was a Mail-Copies-To header in the original, and act accordingly.
Asbjørn Ulsberg
> And if you were to use "followup + reply", it ought to look whether
> there was a Mail-Copies-To header in the original, and act accordingly.
Of course.
Rijk van Geijtenbeek
wrote:
> In article <opr43tfk...@news.opera.com>, Rijk van Geijtenbeek wrote:
>> Thoth is breaking some spec here by adding a space between the new > and
>> the original >, which causes the quote levels to break.
>
> No, it's Opera messing up -- Thoth neither added nor removed any
> spaces, Opera did: "> > Even" is how I sent it,
Yes, and you should have sent it as:
">> Even"
That's what I've been told.
Asbjørn Ulsberg
>> On NNTP? Where? And how?
>
> Oh, not on NNTP, sure. And yes, GZIP would help.
Yes, gzip on NNTP would help considerably. In most cases, you can
get the workload down 50% or even more.
> Nothing to be sorry about. $100 is what extra 0.5-2 GB of traffic
> in Russia will cost you, depending on location.
Oh my god. That's expensive!
> NNTP alone isn't likely to generate so much traffic.
Exactly.
> But it all adds up at the end of the day. No Unicode anywhere is
> better in this sense.
I would still argue that binary files is what fills up the wire, not
text. Maybe the usage patterns are extremely different in Russia, but
that's at least how it is in most other countries I've read network
statistics from (including my own country; Norway).
Unicode might increase the wireload, but I doubt it will be a big
impact on the bigger picture. Alas, I don't understand the problem
with Unicode. That being said, I have no problem with encoding
characters with fewer bits and smaller encoding tables either.
ISO-8859-1 works like a charm in Norway in 99.99% of the cases, and
there's no reason to use anything else at the moment. But whenever
Unicode is needed, it should be available, and it should not cause
any trouble at the recipient side. If it does, the recipient has a
(big) problem, not the sender.
> Sure. The question is: how much is a smilie?
Smilies and other symbols aren't important. What's important for me,
is to stop worrying about what the heck I should encode my data with,
not to mention what encoding I should use to consume other people's
data.
If everyone used Unicode, that wouldn't be something I had to worry
about, and I could focus more of my energy in what's really important
in the application I make; functionality.
> And the bad thing about it is that the recepient can't choose. It's
> exactly the same problem as with HTML mail, IMO. It's an
> embellishment, not a necessity.
I disagree completely with the HTML comparison, but I understand why
you have remonstrances to Unicode. But I'm really glad we have the
standard, and I hope it will be supported everywhere ASAP, so we
can gradually start moving towards an 100% UTF-X world where encoding
is something application programmers (like me) can leave to protocol
programmers to fiddle with.
> I agree 100%. There _are_ occasions when it may be useful, even though
> they are rare. But going back to requirements for newsreaders, I just
> don't see Unicode used anywhere on NNTP, so it's not a big deal anyway.
Not supporting Unicode is halting the progress and evolution. I'm not
saying we should use it recklessly (yet), I'm just saying it's a good
thing to support, so we are better prepared for the future.
Asbjørn Ulsberg
> Just to clarify my position a bit: I don't oppose the idea of Unicode
> as such. I just call on people to be cautious in using it.
I agree with that _now_. All software should support it, but should at
the same time be careful not to use it if not necessary. But tomorrow
(whenever that is), everyone should not only support it, but use it as
if they've never used anything else before.
> Just like many great technologies, I'm afraid it may be easily abused.
> I like the way Opera deals with it, by the way. It's not used by
> default, but can be triggered if needed (and now there's a warning, so
> I can turn it off if I need to). I think _such_ behavior would make a
> good standard.
Absolutely. And there's no reason for the GNKSA (if it were to say
anything about Unicode support) not to reflect this, is it?
Asbjørn Ulsberg
> There is a new NNTP standard in the offing. See
> draft-ietf-nntpext-base-21.txt.
Yes, I actually new that. Stupid of me not to think of it. But I'm
pretty confident the GNKSA will followup (to use NNTP terms) on this
when it is finished and starts getting implemented?
> Indeed. Reusing Reply and Reply-All as Followup and Post-and-Mail is
> just plain sloppiness.
Mhm.
> Likewise using Send when you mean Post.
I agree that the button should say «Post», but as I always use shortcut
keys for just about everything, it isn't really that important to me.
«Reply» / «Followup» is, though.
> (Though might it not be possible to use "Customize toolbars" to fix
> this?).
I have no idea, really.
> What you actually need is a facility to add any arbitrary header to
> any article/message.
Yes, it would work as a temporary solution. But you can't expect
intermediate or rookie users to hack the headers of their articles
just to be able to cancel and supersede them.
It's a great functionality for advanced users, but many permanent
headers can already be set in accounts.ini. Being able to set them
per article would be very nice, though.
> Then people can create the more unusual headers for themselves
> (X-No-Archive anyone?).
I think I've read that Opera doesn't support this intentionally. You
could add all kinds of headers in accounts.ini, but this one will be
ignored, because it's pure vandalism. Want to be anonomous? Don't
post.
> This is where GNKSA needs to catch up with recent developments, like
> format=flowed.
Yes, it should probably mention it.
> And 80 is not necessarily the right number in all circumstances
> (actually, the proper number should be 79 is you want a hard number).
I think I've set up Xnews to use 76 or something. It's below 80 at
least.
> And the Xnews hard break is certainly wrong (for URLs if nothing
> else).
It's not perfect, but I actually prefer it over the total lack of
control I have in Opera right now.
>> Both scoring and filtering is something I miss deeply in M2
(I didn't mean «filtering» here, but «killing»)
> Yes, that is the biggest single failing in M2.
Indeed. The lack of this, and the efficient navigation of Xnews, makes
Opera a much more tedious newsreader, and I find myself using a lot
more time reading news with it, than I did with Xnews. I hope this
improves drastically.
> And filters do not really do the job, because you would need to define
> as many views as you have newsgroups.
Absolutely. I hope «Active threads» can be improved enough to fix much
of this, though.
>> It makes sure that the References-header doesn't exceed 998 characters
>
> Better to fold rather than truncate. And Usefor thinking is to trim when
> it gets to 21 entries (as opposed to GNKSA which says 3).
Mkay. Then Opera may not be 100% conformant, but I believe it's good
enough, no?
> What it wants is a means to turn format-flowed right off for special
> purposes.
In my case, that would be all the time, but yes; I would love to be able
to turn it off.
> What we have is just the ability to turn off or alter the physical line
> length - useful as far as it goes, but not suficient.
I want a visual wrap and a ruler. :-)
J.B. Moreno
> J.B. Moreno <pl...@newsreaders.com> wrote:
> > Rijk van Geijtenbeek wrote:
>
> >> Thoth is breaking some spec here by adding a space between the
> >> new > and the original >, which causes the quote levels to break.
> >
> > No, it's Opera messing up -- Thoth neither added nor removed any
> > spaces, Opera did: "> > Even" is how I sent it,
>
> Yes, and you should have sent it as:
>
> ">> Even"
>
> That's what I've been told.
Someone told you wrong then -- there's no spec or convention that
requires that; the reverse is closer to being true:
More often than not, quoting is done by adding "> " to whatever is
being quoted, regardless of whether it's been quoted or not.
While ">" is also fairly common, when used it is used throughout.
Hmn, I think I may see where your misunderstanding came from -- if
Thoth understood format=flowed AND was converting the quoted text when
responding, /then/ ">" would be used for previously quoted text, but
not necessarily for the newly quoted text.
But Thoth doesn't do format=flowed, and even if it did, it wouldn't
necessarily be the right thing to do to attempt to convert fixed text
to flowed text automatically. A good case can be made that whenever a
format=flowed program responds to a message that's not flowed, it
should used format=fixed instead.
--
J.B. Moreno
J.B. Moreno
> On Tue, 16 Mar 2004 20:23:20 GMT, Charles Lindsey wrote:
>
-snip-
> >> It makes sure that the References-header doesn't exceed 998 characters
> >
> > Better to fold rather than truncate. And Usefor thinking is to trim when
> > it gets to 21 entries (as opposed to GNKSA which says 3).
>
> Mkay. Then Opera may not be 100% conformant, but I believe it's good
> enough, no?
GNKSA allows for 3 id's, but recommends not exceeding the 998 octets.
I.e. it doesn't want you trimming at 3, it considers that the minimum
to maintain good threading.
--
J.B. Moreno
Asbjørn Ulsberg
>> I don't have to rewrap manually in Xnews either.
>
> You do when you edit the quoted text -- i.e. when snipping or inserting
> a comment as I've done here.
It manages the most of it automatically, but of course, if I copy+paste
something into the edit window, I need to hack it myself. But that
doesn't bother me at all, and I do it extremely seldom.
> Well, I obviously can't say whether you'll have to get used to it or
> not, but IMO it's certainly a reasonable wish -- I like knowing how my
> post looks as well.
Mhm.
> I wrote the rewrapping code for Xnews (and then Luu greatly extended
> it), so take my word for it -- it's more error prone.
I believe you!
> I'm not sure if Mozilla does this via CSS or not, but it's plain text
> editor does it for unquoted text (quoted text can be longer).
As Mozilla has its XUL for rendering stuff, I'm sure they do a lot of
stuff a bit more proprietary than with CSS. But I really have no idea.
Asbjørn Ulsberg
> Things went wrong with the quoting in the piece above...
Yes, it's stupid of software to insert spaces between the quoting
characters. Another star to Xnews here, as it recognizes much more
characters (or combinations of them) as quoting characters, and you
can even define your own through regular expressions.
I don't wish Opera necessarily should support user-hacked regex'es,
but it should recognize more quoting characters out of the box than
it does now. I see no reason not to recognize '> > >' as the same
as '>>>'.
> Thoth is breaking some spec here by adding a space between the new >
> and the original >, which causes the quote levels to break.
There are many news clients that inserts space between the quoting
characters, and though this is really bad behavior, I think it's just
as stupid not to recognize this and remove the spaces in well-behaved
clients.
> It wraps the quotes directly (as displayed above, it does better when
> proper > characters are used). If you don't add anything to the quoted
> lines, no change will occur there.
But often I would like to hack the quoted text a bit, and I _don't_
write my answers on one single line -- I break them at ~76 characters,
because that's easier for me to both read and write.
> The breaking of your typed reply at 76 characters? Why do you care,
> unless composing ASCII art?
It's both easier to read and write text that's wrapped at 76 characters,
than text which isn't wrapped at all. Having more than 200 characters
on one line doesn't improve readability. At least not for me.
And ASCII art is also a good reason to allow custom wrapping, and show
this in the edit window.
> .. it would indeed be nice if I could be sure of what I did, and a
> visible helpline would be nice then.
Absolutely.
> CSS doesn't have a length value that counts characters, alas. The values
> 'em' or 'ex' are not really suited for this.
'ex' (the height of the letter 'x') _should_ afaik be suited for this on
monospaced fonts, but it seems nobody really cares about (or everyone
miunderstands) that unit.
> Try to add this to your mime.css:
>
> omf|body, omf|div {max-width: 35em}
Yes, that works pretty well. But it still doesn't help me while writing.
:-)
Asbjørn Ulsberg
>> Who the frick cares about IMAP!? :p
>
> Me :p
Okai!
> (If the poor guy has to rewrite the IMAP again, then I think he should
> know that at least someone appreciates it.)
True. That's very considerate of you. :-)
> For me, it's whatever gets the job done easiest. And despite my usual
> leanings that specialization works better is most cases, it's easier for
> me now to use Opera for email, news, and browsing, than to open Pine
> alongside Opera.
That's the exact same conclusion I've come to regarding Opera+Xnews.
If Opera can do it all, I use Opera for it all. But Opera should have
great news software like Xnews and Pine as a goal, because it isn't
good enough yet.
Asbjørn Ulsberg
> The lines in my signature are exactly 79 characters long (see below).
> When you bring up the Compose window, Opera preloads my signature at the
> bottom. Fine!
Yup.
> But when it sends the message, it observes that the length of the lines
> is longer than 76 characters (or 74 characters in my setup). So it splits
> each line in two before sending it.
Hm. I haven't tested, but maybe it does it with my signature as well,
though
it is only 58 characters long?
> Those who read it in flowed-enabled readers will see it restored to its
> former glory. Those who read it in conventional readers will see it
> chopped
> up.
Mkay. That's bad.
> So, to prevent that happening, I have had to replace all the space
> characters by non-breaking spaces.
I understand. That shouldn't be necessary.
J.B. Moreno
> Rijk van Geijtenbeek wrote:
>
> > Things went wrong with the quoting in the piece above...
>
> Yes, it's stupid of software to insert spaces between the quoting
> characters.
Traditionally, the quoting character is actually TWO characters: "> ".
And in fact it's violating this tradition which has obviously caused a
problem in this case....
Original text: "> > Even ", quoted as ">> > " and then the end of that
line was FLOWED as ">> don't".
So, what happened was the original post (mine) used two characters,
Opera used just one -- but when it uses just one and the next character
is a quote character, it then considers that to mean two levels of
quoting.
> Another star to Xnews here, as it recognizes much more
> characters (or combinations of them) as quoting characters, and you
> can even define your own through regular expressions.
Yeah, that's where Luu ran with what I had -- my code just recognized
combinations of spaces and >.
> I don't wish Opera necessarily should support user-hacked regex'es,
> but it should recognize more quoting characters out of the box than
> it does now. I see no reason not to recognize '> > >' as the same
> as '>>>'.
Opera should either convert the message it is quoting as much as
possible or convert it as /little/ as possible. Recognizing the first
level of quoting in quoted text isn't any safer and is actually *more*
error prone.
--
J.B. Moreno
Rijk van Geijtenbeek
wrote:
> Asbjørn Ulsberg <asbj...@hotmail.com> wrote:
>
>> Rijk van Geijtenbeek wrote:
>>
>> > Things went wrong with the quoting in the piece above...
>>
>> Yes, it's stupid of software to insert spaces between the quoting
>> characters.
>
> Traditionally, the quoting character is actually TWO characters: "> ".
For the first quote, yes. For second level quotes, using ">> " is very
common. And in RFC 2646, it is clearly explained that this is the *only*
proper way. Read section 4.5.
http://www.faqs.org/rfcs/rfc2646.html
Peter Varlien
wrote:
> On Fri, 19 Mar 2004 10:48:15 -0500, Brian D Clary wrote:
>
>>> Who the frick cares about IMAP!? :p
>>
>> Me :p
>
And me!
Btw... Opera 7.23/Linux's IMAP doesn't work very well with Telenor
Mobil's (mobilpost.no) IMAP:
Stuff I mark as read with Opera keeps coming back as unread.
> Okai!
>
>> (If the poor guy has to rewrite the IMAP again, then I think he should
>> know that at least someone appreciates it.)
>
> True. That's very considerate of you. :-)
>
>> For me, it's whatever gets the job done easiest. And despite my usual
>> leanings that specialization works better is most cases, it's easier
>> for me now to use Opera for email, news, and browsing, than to open
>> Pine alongside Opera.
>
> That's the exact same conclusion I've come to regarding Opera+Xnews.
> If Opera can do it all, I use Opera for it all. But Opera should have
> great news software like Xnews and Pine as a goal, because it isn't
> good enough yet.
>
--
Peter Varlien
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/
Charles Lindsey
>Rijk van Geijtenbeek <ri...@opera.com> wrote:
>> Yes, and you should have sent it as:
>>
>> ">> Even"
>>
>> That's what I've been told.
>Someone told you wrong then -- there's no spec or convention that
>requires that; the reverse is closer to being true:
No they didn't! The requirement for it to be '>>' is hard wired into the
spec of format=flowed (see RFC 3656).
The requirement has also been in USEFOR (well, in USEAGE now) since way
back.
van Grieg
<pvar...@NOSPAM.online.no.invalid> wrote:
> On Sun, 21 Mar 2004 10:49:26 +0100, Asbjørn Ulsberg
> <asbj...@hotmail.com> wrote:
>
>> On Fri, 19 Mar 2004 10:48:15 -0500, Brian D Clary wrote:
>>
>>>> Who the frick cares about IMAP!? :p
>>>
>>> Me :p
>>
> And me!
Well, the fact that _I_ care about IMAP is no news I guess. :)
--
van Grieg
J.B. Moreno
> J.B. Moreno <pl...@newsreaders.com> wrote:
>
> > Asbj¯rn Ulsberg <asbj...@hotmail.com> wrote:
> >
> >> Rijk van Geijtenbeek wrote:
> >>
> >> > Things went wrong with the quoting in the piece above...
> >>
> >> Yes, it's stupid of software to insert spaces between the quoting
> >> characters.
> >
> > Traditionally, the quoting character is actually TWO characters: "> ".
>
> For the first quote, yes. For second level quotes, using ">> " is
> very common.
Not really -- you don't see a lot of mixtures, it's either ">" or "> "
for everything.
(BTW - was that extra space the result of a bug in Opera or something
you put in intentionally?)
> And in RFC 2646, it is clearly explained that this is the *only*
> proper way. Read section 4.5.
RFC 2646 specifies how format=flowed is handled; if it's not a
format=flowed message, then it doesn't apply -- and this newsreader
doesn't handle format=flowed.
For the message I posted, the content was absolutely correct, for the
message *you* posted, Opera did the wrong thing -- it could have done
one of two things without causing problems:
1) use "> " for all quoted text and no wrapping of quoted lines
2) convert both my text and the text I quoted to "flowed" (a
generally safe option, if general conventions are followed it'll only
fail when specific line lengths are important).
instead it tried for door number 3:
3) use ">" for text I'd quoted and "> " for my text, and then treat
the the whole thing as flowed.
This didn't work, because my quoted text doesn't satisfy the
requirements for format=flowed text, not surprising since my text
/wasn't/ format=flowed.
While this is hardly the biggest flaw ever seen, it s a flaw on Opera's
part -- one that should eventually be fixed, one way or another (if it
doesn't become moot first as everyone supports format=flowed, but I
don't consider that likely in the short term).
--
J.B. Moreno
J.B. Moreno
> "J.B. Moreno" <pl...@newsreaders.com> writes:
>
> >Rijk van Geijtenbeek <ri...@opera.com> wrote:
>
> >> Yes, and you should have sent it as:
> >>
> >> ">> Even"
> >>
> >> That's what I've been told.
>
> >Someone told you wrong then -- there's no spec or convention that
> >requires that; the reverse is closer to being true:
>
> No they didn't! The requirement for it to be '>>' is hard wired into the
> spec of format=flowed (see RFC 3656).
Well, I made a mistake and left out "that requires that for non-flowed
text", but I don't feel so bad about it because I'm pretty sure you
didn't mean to refer to 3656...
RFC 3656 - The Mailbox Update (MUPDATE) Distributed Mailbox Database
Protocol
It's 2646.
> The requirement has also been in USEFOR (well, in USEAGE now) since way
> back.
If so, it shouldn't -- it should either use one thing throughout, if
it's going to use smarts, it should use some real smarts and use
format=flowed ..
--
J.B. Moreno
Charles Lindsey
>Well, I made a mistake and left out "that requires that for non-flowed
>text", but I don't feel so bad about it because I'm pretty sure you
>didn't mean to refer to 3656...
>It's 2646.
Oops! Actually, it's 3676.
J.B. Moreno
<c...@clerew.man.ac.uk> wrote:
> In <230320041938433621%pl...@newsreaders.com> "J.B. Moreno"
> <pl...@newsreaders.com> writes:
>
> >Well, I made a mistake and left out "that requires that for non-flowed
> >text", but I don't feel so bad about it because I'm pretty sure you
> >didn't mean to refer to 3656...
>
> >It's 2646.
>
> Oops! Actually, it's 3676.
Damn, you mean they've released an update to 2646 without fixing it's
two major problems (no way to say "don't wrap this line" or to indicate
external quoting that should be flowed).
--
J.B. Moreno
Charles Lindsey
>In article <Hv35z...@clerew.man.ac.uk>, Charles Lindsey
>> Oops! Actually, it's 3676.
>Damn, you mean they've released an update to 2646 without fixing it's
>two major problems (no way to say "don't wrap this line" or to indicate
>external quoting that should be flowed).
And no way to have indented, flowed text.
But the whole thing is cunningly designed so that any change whatsoever to
its features will inevitably not be backwards compatible :-( .
J.B. Moreno
> "J.B. Moreno" <pl...@newsreaders.com> writes:
>
> >In article <Hv35z...@clerew.man.ac.uk>, Charles Lindsey
> >> Oops! Actually, it's 3676.
>
> >Damn, you mean they've released an update to 2646 without fixing it's
> >two major problems (no way to say "don't wrap this line" or to indicate
> >external quoting that should be flowed).
>
> And no way to have indented, flowed text.
>
> But the whole thing is cunningly designed so that any change
> whatsoever to its features will inevitably not be backwards
> compatible :-( .
That's somewhat true, but given the frequency of use for the extra
features, that's not really a big problem. Unfortunate, but not IMO a
reason to never upgrade the feature list (and of course, the soonest
started on that, the less damage done).
--
J.B. Moreno
kirk fontaine
messages saying this newsgroups are not valid to reply to
Am I doing something wrong here?
On Tue, 16 Mar 2004 02:07:15 +0100, Asbjørn Ulsberg <asbj...@hotmail.com>
wrote:
> On Tue, 16 Mar 2004 01:30:49 +0100, Rijk van Geijtenbeek wrote:
>
>>> I haven't walket through the form _very_ closely, but am planning
>>> to do the full evaluation some day.
>>
>> It didn't take me that long yesterday :)
>
> Did you submit it as well?
>
>>> As long as the NNTP standard doesn't evolve, there's no apparent
>>> reason why the GNKSA should either, no?
>>
>> The lack of mention of Unicode/UTF8 support and format=flowed is a pity.
>
> True.
>
>> A program like Gravity gets a seal, while it is (I've heard) unsuited
>> for handling UTF-8.
>
> Xnews doesn't handle UTF-8 either, but still has the GNKSA seal. I think
> non-UTF-8 support is OK, but newer clients SHOULD support it.
>
>> People keep asking for hacks to send all kinds of encodings, because
>> so many programs still in use can't handle UTF-8.
>
> Yup. In Norway, the problem is kind of limited, because everyone can just
> default to ISO-8859-1. But in international groups, it becomes a lot more
> complicated, and the easy solution is of course UTF-8.
>
>> And people keep using clients that don't do format=flowed display,
>> which would make their experience so much nicer, and would make the
>> requests for (automatic or manual) reformatting quoted material
>> irrelevant.
>
> I don't fully agree with that. Even with 'format=flowed', I try to make
> my articles pretty, keep the lines within 80 characters, etc. I think
> this is something good, at least that I _can_ control this. I don't think
> I have enough control in Opera, and it makes me very anxious.
>
>> So, I think GNKSA 2.1 MUST demand UTF-8 support and SHOULD encourage
>> format=flowed support.
>
> I agree to UTF-8, but the 'format=flowed' doesn't really mean diddlysquat
> to me. But I don't mind seeing it as a part of the GNKSA.
>
>> The necoiding issue is mentioned on the 'todo' page (unchanged since
>> early 2001). Some submissions are listed as being worked on (also
>> since early 2001).
>
> Maybe we should investigate in how we can get the document updated?
>
>> I think so as well. Let's get out of this IMAP and SMTP business and
>> concentrate on the really important protocol: NNTP :)
>
> Yea! #:-D
>
>> I don't think it is so terrible, and at this moment many people don't
>> use their real email address in postings. Trying to reply by mail only
>> is not the most important feature IMHO.
>
> My problem lies in the fact that news and e-mail isn't clearly enough
> separated in M2. That's what I'm talking about.
>
>>>> 3b Warns about, or prevents, posting to large numbers of groups
>>>
>>> Do you know for sure that Opera doesn't warn about this?
>>
>> No, not for sure. But I've never heard about it, and I haven't seen
>> the error message in the language file for example.
>
> Ok. Let's say it doesn't, then. :-)
>
>>>> 3c Strongly encourages setting of a "Followup-To: " for large
>>>> crossposts (SHOULD)
>>>
>>> Same as above.
>>
>> I'm quite sure there is no UI for that encouragement.
>
> :-\
>
>>> a SHOULD is a SHOULD is a SHOULD. :-)
>>
>> But a SHOULD is not a MUST :)
>
> Not quite, but almost. And for a supreme piece of software like Opera,
> every SHOULD ought to be upgradet to a MUST (now you might ask what
> happens to MUST then? Does it go all the way around the circle and
> end up as a MAY? ;).
>
>> How does Xnews handle large URLs (sending and receiving)?
>
> It cheats. You can select a large peice of text and right-click
> 'Edit URL', edit away all the nastyness and then execute the URL to
> the operating system. It supports <url:..> schemes partly, but the
> 'Edit URL' is the base support Xnews has for URIs.
>
>> I find that, in practice, Opera behaves very well, as long as people
>> don't get confused and start entering their own ENTERs...
>
> Hm. This should work, as long as the URI is enclosed in <url:..>
> brackets.
>
>> I've never figured out how to use Xnews properly in this regard.
>
> The best way to do scoring in Xnews, is to edit the score file and
> manually hack together a regex which matches your criterias. I've
> hacked together a regex which recognizes 'Asbjørn Ulsberg' in the
> 'From' header.
>
> Every post with a 'Reference' set to an article posted by me, also
> gets scored up (and highlighted). Also, the whole thread gets a
> higher score than usual (and thus gets prioritized higher in the
> sort order).
>
>> But some simple 'watch' and 'hide' functionality for both threads
>> and persons is nice. We've got some of it, but not really enough.
>
> I don't think we have anything compared to Xnews. Xnews is just
> brilliant here, although a bit too difficult to set up. If Opera
> could embrace all that's good in Xnews somehow, it would be just
> perfect!
>
>> See above - how often do you reply by mail only?
>
> Seldom, but when I do, I would expect exactly the same as the GNKSA.
>
>> And without having to manually edit the address anyway?
>
> Also seldom, but removing an '.invalid' part of the e-mail address
> is different (and more common) than removing all sorts of other
> headers because M2 doesn't have implemented a way to directly reply
> by e-mail.
>
>>>> [3] The word 'followup' is not used, but 'reply' instead
>>>
>>> Yup, not good.
>>
>> The perils of the integrated interface. I doubt this will be changed
>> anytime soon.
>
> Me too. It's not critical either. Even if there is a technical difference
> between follow-up's and reply's, I don't find it that important. 'Reply'
> works fine, if people just know what they're replying to. Here's where
> I think M2 is too vague.
>
>>> Yup. Not good at all. I hate it, infact.
>>
>> What exactly? Format=flowed in general, or that the compose window
>> doesn't indicate where it will break?
>
> Mostly that the compose window doesn't indicate where it will break. But
> I tend to think 'format=flowed' gives me way too long lines as well, and
> I haven't found a good and simple way to control this in M2 yet.
>
>> With format=flowed, it will not matter when your readers are using
>> capable software (Eudora, Hotmail, Mozilla Mail, Opera, Apple Mail),
>> because their display adapts to the window width, which is much nicer.
>
> True, but it also takes away a bit of the control, I feel. It should
> be optional, at least.
>
>> Unless you are composing ascii art.
>
> Do not underestimate the power of ASCII! I draw in ASCII all the time,
> actually. Attaching images is no-no, and uploading to a web site takes
> way too much time. :-)
>
>>> No, but common sense kind of says it should, no? :-)
>>
>> Right. People pressing 'Reply', and then 'Send' while forgetting to
>> actually enter your own text are also an annoyance, and it would be
>> nice if Opera would warn about that.
>
> Yep. Especially when it's so frikkin' easy! ;-)
>
--
yours truly kirk fontaine personal fitness trainer
www.profile.yahoo.com/kftrainer
kirk fontaine
On Thu, 18 Mar 2004 21:29:06 -0500, J.B. Moreno <pl...@newsreaders.com>
wrote:
> Asbjørn Ulsberg <asbj...@hotmail.com> wrote:
>
>> On Mon, 15 Mar 2004 20:03:43 -0600, Adam Bailey wrote:
>
>> >> 'format=flowed' doesn't really mean diddlysquat to me.
>> >
>> > It'll mean something to you when you can resize a window and have
>> > everything still look good.
>>
>> Neh, not really. I like fixed width in news. I like Courier New as
>> a news and e-mail font as well. I dislike long lines in these mediums.
>
> Even with a fixed width, format=flowed has it's advantages -- you don't
> have to keep manually rewrapping the quoted text, and when snipping,
> things just end up looking good.
>
> Not that you can't approximate this without format=flowed, but it's
> more error prone.
>
> Also, there is nothing in format=flowed that says you can't wrap the
> lines at a fixed width for display -- I think that'd be a good idea as
> a default as I recently told someone else implementing f=f.
kirk fontaine
it as a mail and news client was using Eudora until it starting acting
funny recieving my mail from my email server so I tried opera again and it
work so even with its flaws I deal with it better on my resources I
think
Maybe the final version will be better huh?
On Sun, 21 Mar 2004 10:49:26 +0100, Asbjørn Ulsberg <asbj...@hotmail.com>
wrote:
> On Fri, 19 Mar 2004 10:48:15 -0500, Brian D Clary wrote:
>
>>> Who the frick cares about IMAP!? :p
>>
>> Me :p
>
> Okai!
>
>> (If the poor guy has to rewrite the IMAP again, then I think he should
>> know that at least someone appreciates it.)
>
> True. That's very considerate of you. :-)
>
>> For me, it's whatever gets the job done easiest. And despite my usual
>> leanings that specialization works better is most cases, it's easier
>> for me now to use Opera for email, news, and browsing, than to open
>> Pine alongside Opera.
>
> That's the exact same conclusion I've come to regarding Opera+Xnews.
> If Opera can do it all, I use Opera for it all. But Opera should have
> great news software like Xnews and Pine as a goal, because it isn't
> good enough yet.
>
--
Richard Grevers
wrote:
> what is GNKSA?
>
see http://www.gnksa.org/
--