I'd be quite happy to do this when I bump into links that no longer work, but
I was concerned that people might not like that.
>F.ex. try clicking
>on some of the links in the IRC section for IRC software (I'm still looking
>for the ultimate IRC client).
Oh, easy, IMO... http://sourceforge.net/project/shownotes.php?release_id=105960
but it's still too new and unfinished for a banner "grand opening". Progress is
going slow, though. I just got the windows test GUI taking in full unicode with
IME. Passing it over IRC is no problem using CTCP/2 attributes and a special
"filter" on the way out. The UI half of it needs the most attention.
I used to follow Go-IRC, but I don't ever recall it being Tcl enabled. Maybe in
the very early beginning it was, but has since, completely, disappeared.
I went through a few of the links, but it'll take some time..
--
David Gravereaux <davy...@pobox.com>
Tomasoft Engineering, Hayward, CA
[species: human; planet: earth,milkyway,alpha sector]
I would say that just bluntly removing the references is too easy a way.
Perhaps the following would be better:
- If it is clear to what it was supposed to link to, for instance,
an extension of some sort, replace it with:
[[link to "package" - correct URL unknown]]
- If it is unclear what was meant, then I suppose just cleaning up the
text (removing the reference) is better.
- Add a reference to the page in question on a reserved Wiki page,
called "Links need attention" or something like that.
Just removing the references feels like tearing out the literature
references
because you can not find the books in the library ;-)
Regards,
Arjen
Once again I think that what is missing most in the tcl wiki is a revision
control system known from other wiki software (e.g. http://www.usemod.com).
Without that the community has no real control over the content, I feel.
David
Yes, I looked at that actually but had difficulties getting it to work.
At least with Linux.
For me the most important criteria are:
* Tcl scripting
* No annoying popups all over the place.
* Should have some ircII -style commands
* Preferrably works without a GUI
At least you're actively working on it. I tried dirc and quirc and some
others, but the authors haven't responded to emails on some issues.
In a way, yes. The trouble is we'll then get more and more references to
software which simply does not exist any more.
>David Gravereaux <davy...@pobox.com> wrote:
>> Kristoffer Lawson <se...@gfanrend.fishpool.fi> wrote:
>>
>>>F.ex. try clicking
>>>on some of the links in the IRC section for IRC software (I'm still looking
>>>for the ultimate IRC client).
>>
>> Oh, easy, IMO... http://sourceforge.net/project/shownotes.php?release_id=105960
>
>Yes, I looked at that actually but had difficulties getting it to work.
>At least with Linux.
Yes, no makefile for irc_engine, for one. And not even a Tk UI as of yet. All
my work is in compiled code which is rather backwards :)
>For me the most important criteria are:
>
>* Tcl scripting
Got that. java, python and perl are on the drawing board as well, but Tcl is
the top of the food chain as that's the glue that holds it together.
>* No annoying popups all over the place.
>* Should have some ircII -style commands
>* Preferrably works without a GUI
The main feature this will have is replaceable UIs through a little [Incr Tcl]
inheritance. I've been spending my time on the architecture rather than
sprinting for the finish line. The command API for controlling the UI needs the
most attention right now. I've been distracted getting the unicode support
working.
screenshot:
http://tomasoft.sf.net/doing_unicode.gif
>At least you're actively working on it. I tried dirc and quirc and some
>others, but the authors haven't responded to emails on some issues.
--
Libraries have empoyees to help you, wiki doesn't. ISBN number is a good
thing.
>
> In a way, yes. The trouble is we'll then get more and more references to
> software which simply does not exist any more.
People move a lot, data too. Sometime the software just moved. Then there is
no reason to keep reference to old locations. A problem that will arise is
the use of dynamic binding of IP adresses and these kind of servers as more
and more people get a more or less permanent connection to the net. These
connections get cut every now and then and the server must register itself
again under its name after coming up again. If you, meanwhile, try to access
this page and get a 404 and then prompt delete the link just because you
can't find it, you have removed the presence of an existing software.
Put the links somehow on observation and don't let the search machine put
these kinds of references at the top of a search result. Dead links is a
reason for loosing patience with an information system.
Svenn
What would you leave in place of the links? If the basic information
about the package would be left - author name, description, perhaps original
package name, then people would be able to search google, old CD distributions
etc for a package.
If however, people began making (ala Orwell's 1984) historical adjustments
by removing all references to packages, then I would be less than enthused
about things.
_My_ preference would be if people who are interested would work with
Project E.L.M.O. <URL: http://wiki.tcl.tk/ProjectE.L.M.O. > to work out
some sort of future direction for references in general.
Over the past few years, JM Ivler, Brent Welch, myself and others have
taken on various attempts to categorize the web resources within the
community. Only a certain amount of resources are available on personal
levels to provide such resources - without community contributions, the
goal is likely to be less effective.
My intent on providing resources on the wiki is to try to empower the
community to provide the ability to manage things itself. It is my hope
that we thus move from being throttled by one person's time to be throttled
only by the community's contributed energy.
--
Tcl'2002 Sept 16, 2002, Vancouver, BC http://www.tcl.tk/community/tcl2002/
Even if explicitly stated to the contrary, nothing in this posting
should be construed as representing my employer's opinions.
<URL: mailto:lvi...@yahoo.com > <URL: http://www.purl.org/NET/lvirden/ >
However I did find the wiki a little bewildering as a newbie. Then when I
eventually had a play on the graffiti page I thought 'wow this is what it's
about'.
Glenn
Just had a bit of a brainwave (expect I'm not the first person to think
of this though) ...
How about replacing them with links to the Wayback machine at
http://archive.org ?
For example, this link http://www.cs.vu.nl/%7Eerbruijn/ from the irc
page on the wiki is broken, but this link
http://web.archive.org/web/*/http://www.cs.vu.nl/%7Eerbruijn/ will let
you look at the most recently archived version of it.
In fact, one could even write a tcl script to crawl over the wiki
looking for dead links and, if they are still dead a week later, replace
them with such a link ...
Regards,
Kieran Elby
That is an interesting idea.
Note that the wayback machine is not immune to loss of data as well
(See http://yro.slashdot.org/yro/02/09/24/226253.shtml?tid=153). This
might be less of an impact for technical data, but it can happen.
Still, so far one of the better possibilities of managing dead links.
--
Sincerely,
Andreas Kupries <akup...@shaw.ca>
Developer @ <http://www.activestate.com/>
Private <http://www.purl.org/NET/akupries/>
-------------------------------------------------------------------------------
}
Glenn, perhaps you can help us improve this. Why not start a wiki page
for "How to begin using the Wiki" or some such thing - we can put a link
to it near the top of the Wiki page. Just pick a page title that is
clear, consise, and spelled correctly <wink>.
--
Tcl - The glue of a new generation. <URL: http://wiki.tcl.tk/ >
:> In fact, one could even write a tcl script to crawl over the wiki
:> looking for dead links and, if they are still dead a week later,
:> replace them with such a link ...
:
:That is an interesting idea.
:
:Note that the wayback machine is not immune to loss of data as well
:(See http://yro.slashdot.org/yro/02/09/24/226253.shtml?tid=153). This
:might be less of an impact for technical data, but it can happen.
:Still, so far one of the better possibilities of managing dead links.
First attempt to do this resulted in wayback machine reporting no
hit found for the URL attempted.
Too bad this isn't working out. However, if people want to try
out and see if things are found, then replace the links, that would be
fine for ones where the links are found.
I think that's probably a good idea: concentrating on a good architecture
instead of hacking something quick together. Keep us posted. Especially if
a working system runs on Linux. For the moment I think I'll just use
quirc.
---
/ http://www.fishpool.com/~setok/
Sure. I've often done that, but more often than never it is simply impossible
to find the package anywhere. You might find the odd link, which point
to the site that no longer works. I still think that at some point we
should just remove references to the package altogether. There really is no
point in having a growing list of packages which just do not exist.
>
> My intent on providing resources on the wiki is to try to empower the
> community to provide the ability to manage things itself. It is my hope
> that we thus move from being throttled by one person's time to be throttled
> only by the community's contributed energy.
Yes, of course. I'm just asking about this here before starting to clean
up.
Thanks. Someone either here or in email mentioned one or more of the
web sites providing 'archiving' of web pages, etc. as a possible option
for finding things. I guess the best approach might be to research the
web, including procplace.com and other such archives, then asking here on
clt for leads on archives (Tcl Blast or the recent Tcl/2002 conference CD)
to see if the software exists on someone's disk and if anyone would be
willing to help .
--
Tcl - The glue of a new generation. <URL: http://wiki.tcl.tk/ >