Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Proposed change to bookmarks keywords

34 views
Skip to first unread message

Marco Bonardo

unread,
Jun 22, 2010, 9:55:07 AM6/22/10
to
Till today keywords have been associated to bookmarks, so if you have 2
bookmarks with the same uri you could associate them different keywords.
Or you can associate the same keyword to different bookmarks with
different uris.

This causes some headache and unwanted behavior, keywords are intended
as shortcuts for search engines and sometimes as aliases for uris, If
you type in a keyword that has multiple uris you could get weird results
from the locationbar (or at least not what you are expecting), or you
have to make a choice, that is going against the idea of a shortcut,
converting the keyword to a privileged tag.

Anything that is related to a keyword links to a uri.

Thus in Bug 572030 - "Use a memory cache for keyword to uri
associations" - I'm suggesting to implement this as a 1:1 relation, a
keyword can be associated only to a single uri, and a uri only to a
single keyword. This seems to make sense for the keywords use-case.

The API is not going to change though, the current API has
setKeywordForBookmark(aBookmarkId) while it should really be
setKeywordForBookmarkedURI(aURI). I've not changed it so far for 2 reasons:
- backwards compatibility
- getting an URI would be slower, since would mean having to check if
that uri is bookmarked (the database associates keywords on the
bookmarks table sadly, and it's not easy changeable for downgrade
compatibility). The UI on the other side always knows if the uri is
bookmarked and has the bookmark id.

Due to this, calling setKeywordForBookmark() will remove old
associations (if the keyword was associated to another uri) and create a
new one for all bookmarks with the same uri as the one required.
Creating a bookmark with same uri as one that has a keyword will inherit
the keyword.

I'd like to know if anybody has feedback regarding this behavior change.

Marco

David McRitchie

unread,
Jun 22, 2010, 10:48:11 AM6/22/10
to
"Marco Bonardo" <mak77NO...@supereva.it> wrote in message news:yfOdnQyM0LYjXb3R...@mozilla.org...

You really have no concept of the use (usages) of keywords.

Boris Zbarsky

unread,
Jun 22, 2010, 10:57:09 AM6/22/10
to
On 6/22/10 10:48 AM, David McRitchie wrote:
> You really have no concept of the use (usages) of keywords.

You know, as constructive criticism goes, that's not very useful. Could
you give an example of a usage that would break with Marco's proposed
change? That would be a lot more helpful than what is effectively
kindergarten name-calling.

-Boris

Thomas Stache

unread,
Jun 22, 2010, 11:57:27 AM6/22/10
to
On 22.06.2010 15:55, Marco Bonardo wrote:
> Till today keywords have been associated to bookmarks, so if you have 2
> bookmarks with the same uri you could associate them different keywords.
> Or you can associate the same keyword to different bookmarks with
> different uris.
>
> This causes some headache and unwanted behavior, keywords are intended
> as shortcuts for search engines and sometimes as aliases for uris, If
> you type in a keyword that has multiple uris you could get weird results
> from the locationbar (or at least not what you are expecting), or you
> have to make a choice, that is going against the idea of a shortcut,
> converting the keyword to a privileged tag.


I don't agree with this assessment at all - I enjoy having the same
keyword/shortcut for multiple search engines (search bookmarks, not
search field engines) and making my choice from the Awesomebar results. E.g.
"map New York"
would offer searches for "New York" with either Bing Maps or Google Maps, or
"mxr ABC"
offers MXR searches for various products/repositories...

T.

David McRitchie

unread,
Jun 22, 2010, 3:44:55 PM6/22/10
to
"Marco Bonardo"...

> Till today keywords have been associated to bookmarks, so if you have 2
> bookmarks with the same uri you could associate them different keywords.
> Or you can associate the same keyword to different bookmarks with
> different uris.

That is a true statement, it is what makes a keyword function properly.

> This causes some headache and unwanted behavior, keywords are intended
> as shortcuts for search engines

Disagree, not for search engines, but originally probably for search, but have
many other uses as well. The behavior is wanted in both cases.


> and sometimes as aliases for uris,

Actually that is what they are, an alias for a specific bookmark. A bookmark
can be a url, it can be javascript, it can have substitution for example for a
search, or to direct one with a specific substitution to a specific filename,
it can have multiple urls, it can be combinations of any of those.

You can rearrange parts of bookmark and get exactly the results, and that
would not be considered a duplicate uri.

Anyway an alias is a means of pointing to an object. I have have an alias
that one day points to one object and on another day points to a different
object without having to destroy either object. That's how aliases are used.

> If
> you type in a keyword that has multiple uris you could get weird results
> from the locationbar (or at least not what you are expecting),

Disagree, I get what I expect or I don't get what I expect, if it is not
what I expect or what I want later, I can fix it later. The "AwesomeBar"
makes seeing them at the same time a lot easier now.

>or you
> have to make a choice, that is going against the idea of a shortcut,
> converting the keyword to a privileged tag.

Disagree, keywords existed before tags in Firefox, a keyword
is not a privileged tag, nor is a tag a crippled form of keyword they are
distinctly different. Mistakes may be made, so what. You think you are
going to fix my mistakes by destroying what I have, or what I intend to have.

> Anything that is related to a keyword links to a uri.

Disagree, it may be JavaScript, and if the bookmarks requires substitution
is it a uri before it gets the substitution. (does it really matter)

> Thus in Bug 572030 - "Use a memory cache for keyword to uri
> associations" - I'm suggesting to implement this as a 1:1 relation, a
> keyword can be associated only to a single uri, and a uri only to a
> single keyword. This seems to make sense for the keywords use-case.

This makes sense only if your primary purpose continues to be to destroy keywords.


> The API is not going to change though, the current API has
> setKeywordForBookmark(aBookmarkId) while it should really be
> setKeywordForBookmarkedURI(aURI). I've not changed it so far for 2 reasons:
> - backwards compatibility
> - getting an URI would be slower, since would mean having to check if
> that uri is bookmarked (the database associates keywords on the
> bookmarks table sadly, and it's not easy changeable for downgrade
> compatibility). The UI on the other side always knows if the uri is
> bookmarked and has the bookmark id.

Don't ever change that, and don't break my usage.



> Due to this, calling setKeywordForBookmark() will remove old
> associations (if the keyword was associated to another uri) and create a
> new one for all bookmarks with the same uri as the one required.
> Creating a bookmark with same uri as one that has a keyword will inherit
> the keyword.

This is absolute destruction, duplicate keywords may or may not be intended,
for me the duplication is not really intended as a final result, but is one that can
be fixed easily at a later time, when I decide which choice I really want, and will
modify the other one. This proposed destruction would burn bridges and
destroy my bookmarks. Just as the click on bookmark star originally broke
bookmarks until you could later force it to allow duplicates (after losing valuable
resources).



> I'd like to know if anybody has feedback regarding this behavior change.

I'm afraid this is just going to be more repetitions of further destructions,
and if you had learned to use keywords in the first place we wouldn't be
continually subjected to such wanton destruction.

David McRitchie

unread,
Jun 22, 2010, 4:26:19 PM6/22/10
to
"Boris Zbarsky" <bzba...@mit.edu> wrote in message news:8d-dnZqekq3YUr3R...@mozilla.org...
> kindergarten name-calling. // Boris

I have a bookmark with a keyword of ff:
file:///c:/copiedsite/dmcritchie/firefox/%S
and one for ffp:
http://www.mvps.org/dmcritchie/firefox/%S
and another for kws:
http://www.mvps.org/dmcritchie/firefox/kws.htm

on a test copy is might use ff: kws.htm
on the web copy I might use ffp: kws.htm
or strictly on the test copy, I could use kws:
and those who use my stuff could access website with kws:

I distribute portions of my bookmarks that some friends use, and their
version would resolve to the website usage. This would break that usage
denying them the ability to use one of the keywords as one or the other
of the first two would be destroyed.

And that is only the tip of the iceberg.

Message has been deleted

Justin Dolske

unread,
Jun 22, 2010, 5:21:45 PM6/22/10
to
On 6/22/10 12:44 PM, David McRitchie wrote:

>> I'd like to know if anybody has feedback regarding this behavior change.
>
> I'm afraid this is just going to be more repetitions of further
> destructions,
> and if you had learned to use keywords in the first place we wouldn't be
> continually subjected to such wanton destruction.

Have you considered using another browser such as Chrome, Safari, or
Opera? Perhaps you would be happier with a different product. If you
can't participate in a civil discussion, then I'd suggest you not bother
posting at all.

Justin

Marco Bonardo

unread,
Jun 22, 2010, 5:42:48 PM6/22/10
to
Il 22/06/2010 22:26, David McRitchie ha scritto:
> I have a bookmark with a keyword of ff:
> file:///c:/copiedsite/dmcritchie/firefox/%S
> and one for ffp:
> http://www.mvps.org/dmcritchie/firefox/%S
> and another for kws:
> http://www.mvps.org/dmcritchie/firefox/kws.htm

cool, all of these are different uris and I'm not breaking your use case.

> I distribute portions of my bookmarks that some friends use, and their
> version would resolve to the website usage. This would break that usage
> denying them the ability to use one of the keywords as one or the other
> of the first two would be destroyed.

Can you clarify how you distribute your bookmarks and how first 2 would
be broken? I'm missing a piece there.

Thanks
-m

Marco Bonardo

unread,
Jun 22, 2010, 5:50:55 PM6/22/10
to
Il 22/06/2010 21:44, David McRitchie ha scritto:
> "Marco Bonardo"...

>> This causes some headache and unwanted behavior, keywords are intended
>> as shortcuts for search engines
>
> Disagree, not for search engines, but originally probably for search,
> but have
> many other uses as well. The behavior is wanted in both cases.

right, i meant searches and ended up with a wrong phrase, did not mean
"engines".

>> and sometimes as aliases for uris,
>
> Actually that is what they are, an alias for a specific bookmark.

This is a secondary usage that has come out to address lack of
functionality of locationbar engines. Actually few people need aliases
and not lot of them because locationbars are so good at finding what the
user wants.

> Anyway an alias is a means of pointing to an object. I have have an alias
> that one day points to one object and on another day points to a different
> object without having to destroy either object. That's how aliases are
> used.

so, what's the purpose of your alis that points to 2 different objects?
why not having 2 different aliases?

>> If you type in a keyword that has multiple uris you could get weird
>> results from the locationbar (or at least not what you are expecting),
>
> Disagree, I get what I expect or I don't get what I expect, if it is not
> what I expect or what I want later, I can fix it later. The "AwesomeBar"
> makes seeing them at the same time a lot easier now.

So what? if you tag them you can find them easily. Are you using
keywords to find uris in the locatiobar? That's clearly not what they
are intended for, I hope you agree.

>> or you have to make a choice, that is going against the idea of a
>> shortcut, converting the keyword to a privileged tag.
>
> Disagree, keywords existed before tags in Firefox, a keyword
> is not a privileged tag, nor is a tag a crippled form of keyword they are
> distinctly different. Mistakes may be made, so what. You think you are
> going to fix my mistakes by destroying what I have, or what I intend to
> have.

No, just enforcing the 2 different use cases, an alias (keyword) is a
shortcut an uri, a tag is an information that describes uris.

>> Anything that is related to a keyword links to a uri.
>
> Disagree, it may be JavaScript, and if the bookmarks requires substitution
> is it a uri before it gets the substitution. (does it really matter)

both bookmarklets and uris with substitutions are uris, and are going to
continue work.

> This is absolute destruction, duplicate keywords may or may not be
> intended,
> for me the duplication is not really intended as a final result, but is
> one that can be fixed easily at a later time, when I decide which choice
> I really want

I'm only moving your choice from usage to creation. You suggest that you
should be forced making a choice when you need to use it, I think that's
the worst moment. When you actually use the keyword you need it to work,
not to think and design how it should work, that's creation.

-m

David McRitchie

unread,
Jun 22, 2010, 5:51:22 PM6/22/10
to
"Marco Bonardo" <mak77NO...@supereva.it> wrote in message news:4KqdnWAQitzGs7zR...@mozilla.org...

I would change the file:///c:/copiedsite/dmcritchie/ portion to
http://www.mvps.org/dmcritchie/
and it would be in a flat bookmarks.html type of file. When
the user installs it, your changes would change the keyword of
subsequent uses or of previous uses to the same keyword either
ffp: prevails or ff: prevails, even though the bookmarks might even
in different bookmark folders.

Peter Lairo

unread,
Jun 22, 2010, 6:33:53 PM6/22/10
to

I kindly and respectfully and civilly suggest that you may be perhaps
overreacting just a tad. David's post consisted of an overwhelming
majority of useful and factual explanation of why Marco's suggestion was
undesirable. That took valuable time to write. Perhaps you should give
David a bit of leeway before asking him to leave the project entirely
(use Chrome) or to stop participating in it (no posting).
--
Sincerely, Civilly, Respectfully, Politically Correct, ad absurdum...

Peter Lairo

Bugs I think are important:
https://bugzilla.mozilla.org/show_bug.cgi?id=250539
https://bugzilla.mozilla.org/show_bug.cgi?id=391057
https://bugzilla.mozilla.org/show_bug.cgi?id=436259
https://bugzilla.mozilla.org/show_bug.cgi?id=446444

Islam: http://www.jihadwatch.org/islam101/
Israel: http://www.jewishvirtuallibrary.org/jsource/myths2/
Church of the Flying Spaghetti Monster: http://www.venganza.org/
Anthropogenic Global Warming skepsis: http://tinyurl.com/AGW-Skepsis

Marco Bonardo

unread,
Jun 22, 2010, 6:43:33 PM6/22/10
to
Il 22/06/2010 23:51, David McRitchie ha scritto:
> "Marco Bonardo" <mak77NO...@supereva.it> wrote in message
> news:4KqdnWAQitzGs7zR...@mozilla.org...
>> Il 22/06/2010 22:26, David McRitchie ha scritto:
>>> I have a bookmark with a keyword of ff:
>>> file:///c:/copiedsite/dmcritchie/firefox/%S
>>> and one for ffp:
>>> http://www.mvps.org/dmcritchie/firefox/%S
>>> and another for kws:
>>> http://www.mvps.org/dmcritchie/firefox/kws.htm
>>
>
> I would change the file:///c:/copiedsite/dmcritchie/ portion to
> http://www.mvps.org/dmcritchie/
> and it would be in a flat bookmarks.html type of file. When
> the user installs it, your changes would change the keyword of
> subsequent uses or of previous uses to the same keyword either
> ffp: prevails or ff: prevails

The last created association wins.
Not sure why you want to give your users two keywords poiting to the
same exact uri. Does not make more sense removing the useless one from
the html file? after all you are already editing it.

, even though the bookmarks might even
> in different bookmark folders.

indeed, that's one of the purposes of setting keywords on uris, so
copies of the same bookmarked uri are synced up, you can edit keyword in
any of them and have the change propagated without having to edit all of
them.

-m

Shawn Wilsher

unread,
Jun 22, 2010, 6:51:51 PM6/22/10
to dev-apps...@lists.mozilla.org
On 6/22/2010 3:33 PM, Peter Lairo wrote:
> I kindly and respectfully and civilly suggest that you may be perhaps
> overreacting just a tad. David's post consisted of an overwhelming
> majority of useful and factual explanation of why Marco's suggestion was
> undesirable. That took valuable time to write. Perhaps you should give
> David a bit of leeway before asking him to leave the project entirely
> (use Chrome) or to stop participating in it (no posting).
1) Telling Marco he'd think differently if he "had leared to use
keywords in the first places" comes across at being arrogant at the very
least.
2) Justin did not ask him to leave the project. He asked him to
consider using other products because it might make him happier.
3) Justin did not ask him to stop posting. He asked him to post in a
more civil manor (see point one) or to not post at all because it's not
constructive.

Perhaps you should give others a bit of leeway before suggesting they
are overreacting. Justin wasn't the first person to comment on the
non-constructiveness of his comments.

/sdwilsh

Peter Lairo

unread,
Jun 22, 2010, 7:30:23 PM6/22/10
to
On Wed. 23.06.2010 0:51, Shawn Wilsher wrote:
> On 6/22/2010 3:33 PM, Peter Lairo wrote:
>> I kindly and respectfully and civilly suggest that you may be perhaps
>> overreacting just a tad. David's post consisted of an overwhelming
>> majority of useful and factual explanation of why Marco's suggestion was
>> undesirable. That took valuable time to write. Perhaps you should give
>> David a bit of leeway before asking him to leave the project entirely
>> (use Chrome) or to stop participating in it (no posting).
> 1) Telling Marco he'd think differently if he "had leared to use
> keywords in the first places" comes across at being arrogant at the very
> least.

Agreed. But that was a single comment at the very end of a very long and
detailed and rational and civil presentation of David's viewpoint.

> 2) Justin did not ask him to leave the project. He asked him to consider
> using other products because it might make him happier.
> 3) Justin did not ask him to stop posting. He asked him to post in a
> more civil manor (see point one) or to not post at all because it's not
> constructive.

Points 2 & 3 are bending the space-time continuum a bit too much for my
taste. Seems more like "circling-the-wagons" than an objective
description of what was said.

> Perhaps you should give others a bit of leeway before suggesting they
> are overreacting. Justin wasn't the first person to comment on the
> non-constructiveness of his comments.

Just because multiple people share an opinion doesn't make it valid
(see: belief in a god; election of Obama|Bush; number of AOL users > 1)

David's first post was nonconstructive; I fully agree. But the post
Justin repoded to seemed to be an honest attempt (with an imperfect bit
at the very end) to explain in detail why David thought Marco's plan was
undesirable. I don't think Justin needed to pounce on it.

Anyhow. No big deal. I felt Justin was being harsh. You think he wasn't.
Oh well.
--
Regards,

David McRitchie

unread,
Jun 23, 2010, 12:27:22 AM6/23/10
to

"Marco Bonardo" <mak77NO...@supereva.it> wrote in message news:rZmdndrP_qoHobzR...@mozilla.org...


> Il 22/06/2010 23:51, David McRitchie ha scritto:
>> "Marco Bonardo" <mak77NO...@supereva.it> wrote in message
>> news:4KqdnWAQitzGs7zR...@mozilla.org...
>>> Il 22/06/2010 22:26, David McRitchie ha scritto:
>>>> I have a bookmark with a keyword of ff:
>>>> file:///c:/copiedsite/dmcritchie/firefox/%S
>>>> and one for ffp:
>>>> http://www.mvps.org/dmcritchie/firefox/%S
>>>> and another for kws:
>>>> http://www.mvps.org/dmcritchie/firefox/kws.htm

That last one should have been
file:///c:/copiedsite/dmcritchie/firefox/kws.htm
and when sent out for use by others it
would get changed to
http://www.mvps.org/dmcritchie/firefox/kws.htm

>>
>> I would change the file:///c:/copiedsite/dmcritchie/ portion to
>> http://www.mvps.org/dmcritchie/
>> and it would be in a flat bookmarks.html type of file. When
>> the user installs it, your changes would change the keyword of
>> subsequent uses or of previous uses to the same keyword either
>> ffp: prevails or ff: prevails
>
> The last created association wins.

> Not sure why you want to give your users two keywords pointing to the

> same exact uri. Does not make more sense removing the useless one from
> the html file? after all you are already editing it.

No it does not make sense, because you have destroyed one the primarily
aspects of aliases, the ability to have more than one alias.



>> , even though the bookmarks might even
>> in different bookmark folders.
>
> indeed, that's one of the purposes of setting keywords on uris, so
> copies of the same bookmarked uri are synced up, you can edit keyword in
> any of them and have the change propagated without having to edit all of

> them. // marco

But I don't want to change them so that only one keyword can be used that
defeats the who purpose of aliases.

As far as editing goes, the only editing that I am doing is to change
one string to another I am certainly not going to be removing individual
bookmarks as that would be time consuming and would destroy the purpose
of the aliases.

Marco Bonardo

unread,
Jun 23, 2010, 5:51:03 AM6/23/10
to
Il 23/06/2010 06:27, David McRitchie ha scritto:

> "Marco Bonardo" <mak77NO...@supereva.it> wrote in message

>> The last created association wins.
>> Not sure why you want to give your users two keywords pointing to the
>> same exact uri. Does not make more sense removing the useless one from
>> the html file? after all you are already editing it.
>
> No it does not make sense, because you have destroyed one the primarily
> aspects of aliases, the ability to have more than one alias.

This is not an argument, please give arguments why removing the useless
entry from the file instead of editing it does not work for you.

> As far as editing goes, the only editing that I am doing is to change
> one string to another I am certainly not going to be removing individual
> bookmarks as that would be time consuming and would destroy the purpose
> of the aliases.

You are editing a string in an html file, i'm suggesting to remove a
string from an html file, where is the time consuming part?

Provided I understand you are against the change because you think
multiple keywords to the same uri make sense, I'm trying to give you a
solution for your edge use case, let's try to be constructive please.

David McRitchie

unread,
Jun 23, 2010, 12:26:59 PM6/23/10
to
"Marco Bonardo"...

> Il 23/06/2010 06:27, David McRitchie ha scritto:
>> "Marco Bonardo" <mak77NO...@supereva.it> wrote in message
>>> The last created association wins.
>>> Not sure why you want to give your users two keywords pointing to the
>>> same exact uri. Does not make more sense removing the useless one from
>>> the html file? after all you are already editing it.
>>
>> No it does not make sense, because you have destroyed one of the primarily

>> aspects of aliases, the ability to have more than one alias.
>
> This is not an argument, please give arguments why removing the useless
> entry from the file instead of editing it does not work for you.
>
>> As far as editing goes, the only editing that I am doing is to change
>> one string to another I am certainly not going to be removing individual
>> bookmarks as that would be time consuming and would destroy the purpose
>> of the aliases.
>
> You are editing a string in an html file, i'm suggesting to remove a
> string from an html file, where is the time consuming part?
>
> Provided I understand you are against the change because you think
> multiple keywords to the same uri make sense, I'm trying to give you a
> solution for your edge use case, let's try to be constructive please.

I am against the change because it is going to make a bookmark useless.
My bookmarks are not useless to me.
I already indicated what I do, and the reason should be apparent, obviously
it is not because you keep repeating the words "useless". I know pretty well
what I can do or not do with what I have, you would make my examples and
my bookmarks indicated unusable as a means of instruction and of usage,
by your changes.

Rather than telling me bad options about how I should mess up my
bookmarks and those who choose to use my bookmarks, let me tell you
what I would have to do to my bookmarks if your change goes through.

Rather than bookmarking a functional bookmark, you would have me
bookmark instead the uri, this gives you the ability to combine several
keywords to a uri, which sounds good but it really complicates what is
currently simple as currently the keyword resides in the main bookmark
record.

Bad aspects of your proposal and where I think it is going.
1) makes things more complicated for users
2) makes things more complicated for developers, that have
to work with it later.
3) makes it so aliases to the function cannot be easily reversed to
a different uri.
4) makes it difficult to export/import bookmarks as code has to be
rewritten, and be compatible with other versions of Firefox and
with other browsers.
5) makes it impossible to combine segments of bookmarks from
different sources, which is a goal (I think it still is), because there
would be duplicates, and conflicts that have to be resolved, and
removing duplicates by simply removing them is not a viable solution

I have a functional bookmark named
FF: Firefox on local HD
and another named
ffp: Firefox customizations (Notes) (Production)

On my system they point to two different locations, on another
profile or someone else's system they would both probably point to the
production version (on the web). So I would have to put the test version
shortcut name into the production version, and that is probably the preferred
name/title by others unless they had a shortcut name with the same as mine,
and would have to do something different to fit their own needs.
I trust you will be providing for additional aliases (not be be confused with tags)
with your new changes. Of course I would no longer see both names/titles so I've lost
functionality as well for both me and for showing other users how to do something
very simply with keyword shortcuts.

David McRitchie

unread,
Jun 23, 2010, 1:03:54 PM6/23/10
to
Marco Bonardo" <mak77NO...@supereva.it> wrote in message news:9c6dncFoKJnZrbzR...@mozilla.org...

> Il 22/06/2010 21:44, David McRitchie ha scritto:
>> "Marco Bonardo"...
>>> This causes some headache and unwanted behavior, keywords are intended
>>> as shortcuts for search engines
>>
>> Disagree, not for search engines, but originally probably for search,
>> but have
>> many other uses as well. The behavior is wanted in both cases.
>
> right, i meant searches and ended up with a wrong phrase, did not mean
> "engines".
>
>>> and sometimes as aliases for uris,
>>
>> Actually that is what they are, an alias for a specific bookmark.
>
> This is a secondary usage that has come out to address lack of
> functionality of locationbar engines. Actually few people need aliases
> and not lot of them because locationbars are so good at finding what the
> user wants.

That is a false statement, and that was proven false when Alex actually removed
keyword shortcut from Firefox beta. The fact that beta users complained had no
effect but the other developers could not be ignored and it was restored within
24 hours -- very unusual, it also taught Alex something about keyword shortcuts
big change in his blog, or do you disagree with that?

>> Anyway an alias is a means of pointing to an object. I have have an alias
>> that one day points to one object and on another day points to a different
>> object without having to destroy either object. That's how aliases are
>> used.
>

> so, what's the purpose of your alias that points to 2 different objects?

> why not having 2 different aliases?

You did not read what I posted.
Two aliases on my system, one points to test one to production.
Two aliases on another system, both point to production.
Same aliases get to be used on both systems but point to what
is available or to be tested.



>>> If you type in a keyword that has multiple uris you could get weird
>>> results from the locationbar (or at least not what you are expecting),
>>
>> Disagree, I get what I expect or I don't get what I expect, if it is not
>> what I expect or what I want later, I can fix it later. The "AwesomeBar"
>> makes seeing them at the same time a lot easier now.
>
> So what? if you tag them you can find them easily. Are you using

> keywords to find uris in the locationbar? That's clearly not what they

> are intended for, I hope you agree.

Fixing it later means changing the keyword so it is unique, and I can do
that after a couple of weeks after I see which I use the most.

Anything involving tags, I disagree with, very few people are capable of using them
properly on a small system. I am using all keywords directly, only in the
case of forgetting a keyword or if I had a duplicate (I have no duplicates)
keyword I would be interested in looking at the location bar drop-down before hitting
Enter, and I would know what I am dealing with because I include
the keyword in the bookmark name, and always have.



>>> or you have to make a choice, that is going against the idea of a
>>> shortcut, converting the keyword to a privileged tag.
>>
>> Disagree, keywords existed before tags in Firefox, a keyword
>> is not a privileged tag, nor is a tag a crippled form of keyword they are
>> distinctly different. Mistakes may be made, so what. You think you are
>> going to fix my mistakes by destroying what I have, or what I intend to
>> have.
>
> No, just enforcing the 2 different use cases, an alias (keyword) is a
> shortcut an uri, a tag is an information that describes uris.

Actually the description and name is what describes the uri,
if you are using tags properly they would classify the bookmark,
not describe the bookmark. Folders in hierarchy are better suited
for real classification, but modifying the name of the bookmark itself
makes it work very well with the AwesomeBar, much better than tags.



>>> Anything that is related to a keyword links to a uri.
>>
>> Disagree, it may be JavaScript, and if the bookmarks requires substitution
>> is it a uri before it gets the substitution. (does it really matter)
>
> both bookmarklets and uris with substitutions are uris, and are going to
> continue work.

Just a matter of terminology, I know they are going to continue to work,
not afraid of that.



>> This is absolute destruction, duplicate keywords may or may not be
>> intended,
>> for me the duplication is not really intended as a final result, but is
>> one that can be fixed easily at a later time, when I decide which choice
>> I really want
>
> I'm only moving your choice from usage to creation. You suggest that you
> should be forced making a choice when you need to use it, I think that's
> the worst moment. When you actually use the keyword you need it to work,
> not to think and design how it should work, that's creation.

Not at all what I said. I do not have duplicate keywords so forget about that,
it was only mentioned as to how they would be resolved if I had any.

> -m

Alex Faaborg

unread,
Jun 23, 2010, 1:22:25 PM6/23/10
to David McRitchie, dev-apps...@lists.mozilla.org
>
> it also taught Alex something about keyword shortcuts
>

If I remove a feature used by 1% of our user base, I will have roughly 4
million people mad at me. This fact can be interpreted in two different
ways:

-1% is a pretty small fraction of the Firefox user base
-4 million people is a freaking lot of people, it's like the population of
Norway

There's definitely a lesson in here, but I'm not sure we all agree on what
it actually is.

-Alex

> _______________________________________________
> dev-apps-firefox mailing list
> dev-apps...@lists.mozilla.org
> https://lists.mozilla.org/listinfo/dev-apps-firefox
>

Marco Bonardo

unread,
Jun 23, 2010, 2:11:42 PM6/23/10
to
Il 23/06/2010 19:03, David McRitchie ha scritto:
> Marco Bonardo" <mak77NO...@supereva.it> wrote in message

>> This is a secondary usage that has come out to address lack of
>> functionality of locationbar engines. Actually few people need aliases
>> and not lot of them because locationbars are so good at finding what
>> the user wants.
>
> That is a false statement, and that was proven false when Alex actually
> removed
> keyword shortcut from Firefox beta. The fact that beta users complained
> had no
> effect

If you think users complains don't have effects, you're on the wrong way
and you're insulting work of hundreds of persons.
Keywords were and are inteded as search shortcuts, there is no
discussion over that. The fact you can use a car as your house does not
make it an house. Nor it gives us rights to kick you out of it.
The only browser using keywords as full aliases is Opera, indeed they
call the feature Aliases, we also don't show keywords field by default
unless you choice "Add keyword for this search" or you open advanced
properties.
That said, we will never remove the possibility to use keywords as
aliases, what I want to do is to alias an uri, not a bookmark.

>> so, what's the purpose of your alias that points to 2 different
>> objects? why not having 2 different aliases?
>
> You did not read what I posted.
> Two aliases on my system, one points to test one to production.
> Two aliases on another system, both point to production.
> Same aliases get to be used on both systems but point to what
> is available or to be tested.

if the same keyword is used in both system why do you push 2 keywords to
your users instead of one? Would not be easier for him to have 1 simple
keyword too instead of guessing which of the 2 he should use?

> if I had a duplicate (I have no duplicates)
> keyword I would be interested in looking at the location bar drop-down
> before hitting
> Enter, and I would know what I am dealing with because I include
> the keyword in the bookmark name, and always have.

This is not going to change if you don't have multiple keywords pointing
to same uri (or viceversa), you can still do all of this.

>> I'm only moving your choice from usage to creation. You suggest that
>> you should be forced making a choice when you need to use it, I think
>> that's the worst moment. When you actually use the keyword you need it
>> to work, not to think and design how it should work, that's creation.
>
> Not at all what I said. I do not have duplicate keywords so forget about
> that,
> it was only mentioned as to how they would be resolved if I had any.

Are we talking about real use-cases or ipothesis? because if it's the
latter I fear we are losing our time.

-m

Marco Bonardo

unread,
Jun 23, 2010, 2:23:36 PM6/23/10
to
Il 23/06/2010 18:26, David McRitchie ha scritto:
> "Marco Bonardo"...
>> Il 23/06/2010 06:27, David McRitchie ha scritto:
> Bad aspects of your proposal and where I think it is going.
> 1) makes things more complicated for users

Users won't have anymore to choice which uri is the correct one for a
keyword, nor to sync multiple bookmarks to the same uri. once they open
a bookmkark to that uri they can edit keyword and all bookmarks to the
same exact uri will inherit it.
If they assign a keyword of a bookmark that is already assigned to
another uri, they won't find themselves with having to choice and find
the old bookmark to remove the bogus keyword.
This sounds like easier.

> 2) makes things more complicated for developers, that have
> to work with it later.

I can't see an use-case. Even in your use-case you were editing an html
file and after the change you will have to edit an html file.

> 3) makes it so aliases to the function cannot be easily reversed to a
> different uri.

just assign the alias to a different uri.

> 4) makes it difficult to export/import bookmarks as code has to be
> rewritten, and be compatible with other versions of Firefox and
> with other browsers.

what? no, each bookmark pointing to the given uri will have the keyword,
and we manage all of that transparently, nothing changes for
import/export, at all.

> 5) makes it impossible to combine segments of bookmarks from
> different sources, which is a goal (I think it still is), because there
> would be duplicates, and conflicts that have to be resolved, and
> removing duplicates by simply removing them is not a viable solution

This is an edge case where the 2 sources define different keywords for
the same uri or different uris for the same keywords, looks to me
sources are conflicting, so there is no good choice.

> I have a functional bookmark named
> FF: Firefox on local HD
> and another named
> ffp: Firefox customizations (Notes) (Production)
>
> On my system they point to two different locations, on another
> profile or someone else's system they would both probably point to the
> production version (on the web). So I would have to put the test version
> shortcut name into the production version, and that is probably the
> preferred
> name/title by others unless they had a shortcut name with the same as mine,
> and would have to do something different to fit their own needs.
> I trust you will be providing for additional aliases (not be be confused
> with tags)
> with your new changes. Of course I would no longer see both names/titles
> so I've lost functionality as well for both me and for showing other
> users how to do something
> very simply with keyword shortcuts.

So, you have 2 keywords pointing to 2 different uris. nothing changes
here, will work fine.
Then you export them to pass them to a user. after export you edit html
file and change the url of the local resource to the url of the remote
resource.
With my change on import the last set keyword would win (either ff, or ffp).
What i'm suggesting is that instead of editing uri, you remove the duped
bookmark the user is not going to use, or just rename both keyword to
the one he should use.

Let's be clear, I'm not saying you are wrong, but I'm trying to
understand how much harder would be to support your use-case in the new
shape.

-m

Marco Bonardo

unread,
Jun 24, 2010, 8:04:35 PM6/24/10
to
Il 24/06/2010 09:56, Thomas Stache ha scritto:
>> I don't agree with this assessment at all - I enjoy having the same
>> keyword/shortcut for multiple search engines (search bookmarks, not
>> search field engines) and making my choice from the Awesomebar results.
>> E.g.
>> "map New York"
>> would offer searches for "New York" with either Bing Maps or Google
>> Maps, or
>> "mxr ABC"
>> offers MXR searches for various products/repositories...
>>
>> T.
>
> Marco, could you comment on the effect of the proposed change on the
> usage scenario above?

Hi, sorry for late, was working on other fancy stuff.
I find your use-case interesting, will think about it, I got a couple
ideas in the meantime, that could allow me to reach my result without
touching current behavior.
Your use case could be easily replaced by 2 keywords clearly, like
"bmap" and "gmap", but since we still don't have actions, opportunity to
use keywords as configurable-actions sounds interesting.

-m

dmcri...@gmail.com

unread,
Oct 6, 2015, 9:45:12 AM10/6/15
to
On Thursday, June 24, 2010 at 8:04:35 PM UTC-4, Marco Bonardo wrote:
> Il 24/06/2010 09:56, Thomas Stache ha scritto:
> Your use case could be easily replaced by 2 keywords clearly, like
> "bmap" and "gmap",

No you can't because a bookmark keyword is an alias (1:1), and assigned to a specific bookmark. What you are referring to are tags(1:many), and tags are not aliases. Anyway since this went through the remedy is upon the user to know that they already have a bookmark with the same uri content and must make the additional bookmarks unique such as by 1) including a /*comment*/ in their javascript code, or 2) adding/modifying a real or fictitious fragment-id (#alt2) or ? suffixed argument meaningful only to servers and most visibly seen in search engine usage.

> but since we still don't have actions, opportunity to use keywords as configurable-actions sounds interesting.

No ideal what that means, but %S and %s substitutions is what make Firefox bookmarks powerful and configurable especially when combined with Javascript.

dmcri...@gmail.com

unread,
Oct 8, 2015, 11:56:40 AM10/8/15
to

My last sentence should have been expanded concerning power of Firefox bookmarks.


No ideal what that means, but %S and %s substitutions is what make Firefox bookmarks powerful and configurable especially when combined with Javascript, and invoked with a bookmark through it's keyword shortcut (alias).
0 new messages