User talk:SunirShah

From Meta, a Wikimedia project coordination wiki
Jump to: navigation, search

Note to reader. This page was moved to meta.wikpedia.com from somewhere else (probably MeatBall, the original wiki-meta-discussion-site), and while most InterWiki links where fixed to point to something sensible again, because of differing LinkPatterns, some originally intended links might have been lost. So please feel free to do further research in these wikis:


Issues:


As for the TwinPages and FOLDOC, I'm afraid I don't understand. The content on MeatBall:TwinPages wasn't very clear to me, but it seems to be a scheme for synchronizing pages between wikis. If this is correct, then this would have no bearing on our use of FOLDOC content. We're not interested in mirroring FOLDOC, but using content from it to as a base to build our own articles. I would like to hear some more details though, in case I have things wrong. --STG

I'm kind of wondering why Wikipedia is no longer being indexed by MetaWiki. I thought that was pretty cool, anyway. Wouldn't MetaWiki users find it useful? --Larry Sanger


There are more than one MetaWiki. You'll find more on MeatBall:MetaWiki. I suspect Wikipedia stopped listing because http://www.wikipedia.com/wiki.cgi?action=index no longer functions. The best solution to this is to provide a plaintext listing of all the page titles, but I don't have time to develop this myself these days, so that's why I just dropped it. However, now that I discovered the FOLDOC thing, perhaps we should do it.

action=index works just fine on Wikipedia. You just have to give it a minute or so. Probably your script timed out. --AxelBoldt
I just tried it, and it does indeed work... if you have about ten minutes to spare. UseModWiki doesn't seem to be able to handle our 27000+ pages very well. I got to the "L" pages before stoping it; almost 1 meg for half the index. I'm rather disappointed that we were delisted from MeatBall:MetaWiki, without any questions as to what the problem was. A quick question to Jimbo Wales and a test of http://www.wikipedia.com/wiki.cgi?action=index would have cleared up the problem. --STG
Aye, I just delisted it two days ago. Like I said, I just came here to find out what was going on. This page may have started off about TheCuncator reversing a change I made to correct a poor usage of 'Wiki:WikiWiki', and some resolution of the "move chaff onto other wikis" comment, but now we're on the ball. No worries. --ss

This would allow using TwinPages between Wikipedia and FOLDOC. All that means is that whenever two pages of the same name exist on the different sites, a link is put at the bottom to the other site. This is not unlike Google placing a link to dictionary.com at the top of your query. There are several advantages to this scheme.

  • No duplicated effort.
  • More focused sites.
  • Better scaling into non-related domains, such as atlases.
  • Ability to benefit from non-GDLed sites (even though not necessarily those who don't have similar licenses)

An example of where this would have been useful is FOLDOC's incorporation of The Jargon File. As Raymond periodically updates TJF, FOLDOC periodically goes out of sync. Similarly, if Wikipedia updates an entry in FOLDOC, FOLDOC will go out of sync. Possibly also TJF.

I have to check my e-mail at home. I'll do that and get back to you on this. -- SunirShah


Well, Mr. Howe hasn't replied yet, which means that anything we do could only be temporary or experimental. I suggest the following incremental steps:

  • Creation of a plaintext listing of page titles that can be acquired by MetaWiki once a night. This can entire be statically created once a day or be dynamically generated. Minimally you could do a ls -1R on the wiki database.
  • Add a sister pages patch to UseModWiki so it behaves much like Wiki:WikiWiki. That is, pages that don't already exist in the system but exist on sister sites will form a link anyway.

-- SunirShah


I have no problem with proposals a downloadable list of page titles, or you indexing Wikipedia. However, I'm not too sure I like TwinPages or sister pages on Wikipedia. Wikipedia has a different culture from most wikis, and doesn't really fit in with the wider wiki community. How would TwinPages or sister pages help Wikipedia as an encyclopedia?

Finally, we aren't going to stick with UseModWiki. We have our own particular needs as a wiki, and we need custom software to meet those needs. (And personally I have some plans to add non-wiki features to Wikipedia software, such as a database backed yearbook or biographical listing, although whether we'll actually do it I don't know.) -- SJK

I'm not being clear. I don't intend the TwinPages links to be between WikiPedia and, say, Wiki:WikiWiki. The links are to be between WikiPedia and FOLDOC. It's also secondary whether Wikipedia is a wiki or not. http://advogato.org incorporates the InterWiki space through the <wiki/> tag.

To address your second concern, it's irrelevant whether you will be using UseModWiki later. The technology is trivial to implement, so it shouldn't be difficult to move. Since currently you are using UseModWiki, it's possible to upgrade UseModWiki. We should've written a TwinPages patch available for UseModWiki a long time ago anyway, so it's not much loss to do this. -- SunirShah


Sunir,

I found this discussion via RecentChanges, and I read it with great interest. I just wanted to add another voice to the crowd about Wikipedia being a "different" kind of wiki. We're becoming a very focused community rallying around the grand vision of a copyleft encyclopedia. What attracts me most to this project and makes me more active here than on any other wiki is the goal of becoming an authoritative source for information about the material named by the pages. I started out at Ward's wiki , but I lost interest because the community wasn't universal. I've dabbled in the MeatballWiki, but it, too, isn't universal. There's something missing about writing about a topic until there exists a wiki page that steps up and takes on the role of making authoritative and universal statements about that topic. That's why I find Everything2 so unsatisfying that I have never even attempted to edit or create a node there. Wikipedia fills a void on the Web and in wikispace by serving as a repository for a neutral point of view treatment of ... well, of everything! (Ok, so I'm exaggerating. As Larry has pointed out, not everything is suitable for expansion into an encyclopedia article. But there certainly are a lot of pages at Wikipedia yet to be created that are suited to treatment in an encyclopedia.)

About UseModWiki. I have some regret about leaving that software behind, but I agree with the concept of a wiki designed especially for the purposes of this project. (Thanks, Magnus!)

About the confusion over the term "metawiki." We talked about this on Wikipedia-L. See [1], for instance. We're calling http://meta.wikipedia.com/ the "metapedia."

About TwinPages. We really like free links at Wikipedia, and MeatBall:TwinPages says bumpy case is the only proposed support on the table right now. Won't that limit its effectiveness for Wikipedia pages? But assuming free links are eventually supported by TwinPages, I think TwinPages has a chance of working its way into Wikipedia at a later stage in the project. That is, once we have established a reputation as an encyclopedia and not just a wannabe project trying to build an encyclopedia, I've already said that I think we'll have to have a way to represent article content as "released." I see the project moving to a dual state where it has a reputation as a reference for information and it's still under development by Wikipedians (as it always will be). In that state, I think TwinPages makes sense as a tool for Wikipedians -- not so much for readers looking for authoritative information about a topic. I suppose until that time, it would makes sense because it might attract authors from the other wikis to contribute to the encyclopedia article.

Finally, it seems to me that the discussion here is exactly why we are trying to move content about Wikipedia to the metapedia. Wikipedians want to focus on recent changes to encyclopedia articles. I think there's consensus that the metapedia will operate much more like a traditional wiki. <>< tbc

Technical issue first: I invented MeatBall:WikiNameCanonicalization for use by MetaWiki's TwinPages interface. That is, arbitrary typographic changes to a link name all map to the same page with some "canonicalization" of that name. For example, http://sunir.org/apps/meta.pl?wiki=Foldoc&twin=airplanerule. And before anyone complains that such and such a case is not handled, the canonicalization rules are easy to change.

I think UseModWiki does this too, or will be doing this soon, or at least should be doing this. I suggested doing so about a million times.

I agree that this is meta and could legitimately be sloughed onto metapedia. This page kind of evolved here. If someone wants to do what's done to move this page to metapedia, go ahead. I'm not about to attempt to break yet another Wikipedia convention by acting, as I seem to be prone to doing. :)

Regarding your feeling on the timeframe of the change. The point is that currently Wikipedia is in the process of forking FOLDOC. If you think this is a bad thing (and I definitely do), you need another solution. Future considerations are side benefits. If you're interested in thinking about the future, there are potentially many benefits. But I'm not into potentiality. At some point you have to decide Wikipedia is ready. If you're going to decide in favour, it's better to do this before forking a large body of work.

I don't really believe that this will be much more than a fork. If you consider a random page, 10base2 compared to http://foldoc.doc.ic.ac.uk/foldoc/foldoc.cgi?10base2, they are identical. It's not really likely anyone is going to update this page. Or at least the majority of the pages. I find that worrisome, as it's a waste of resources. It's also not how the Internet (used to) works. The nice feature of the 'net is that it's distributed and networked, so the responsibility for resource doesn't have to (nor shouldn't) belong to one domain, but shared amongst many parties. -- SunirShah


I am not at all opposed to Sunir's solution--where FOLDOC has an article about X, and Wikipedia has an article also precisely named X, then exchange links. This doesn't do much harm, and prevents a fork--nearly always a good thing. Besides, it gets people who want to work on FOLDOC to work on Wikipedia!  :-) The trouble, of course, is finding someone to do the necessary coding.  :-(

On the more general issue, bear in mind that it's our goal with Wikipedia to create a huge general encyclopedia. Not a huge general encyclopedia, minus information about computer science (for instance). Other websites don't share this goal or are generally incompatible with Wikipedia's format or aim (e.g., they're proprietary, or they include a lot of, indeed, "cruft"). By creating a single repository of all knowledge (at some general level), as general encyclopedias aspire to be (after all!), Wikipedia surely will not render the Internet any less distributed than it is right now.

Indeed, one might take your reasoning to its logical conclusion, Sunir, and say that the WWW is already a huge distributed encyclopedia. And so it is, no doubt, much more useful than Wikipedia is at present, and given that Wikipedia is part of the WWW and that the WWW will always be much larger than Wikipedia (and that size implies usefulness--hmmm), the web will always be more useful than Wikipedia. But look at it like this--part of what makes Wikipedia so cool is that all this attention and energy is focused on creating a neutral, encyclopedic resource about everything, and everyone working on every part of it has (or should have) more or less the same goal, roughly (!) the same standards, and similar understanding of the most general Wikipedia policies. It's difficult to make that distributed. Alternatively, one could consider the efforts of the 100+ people per day who work on Wikipedia to be a greatly distributed effort!

(Actually, maybe this page should live on its own discussion page here on the "metapedia." Maybe we shouldn't use the name "metapedia," by the way: it's registered. http://www.metapedia.com ) --Larry_Sanger


Hmm, I don't think the move to meta.wikipedia.com was a good idea. The syntax processing isn't working correctly, and everytime I load a page, my browser gives me an error. That will make this discussion more painful than it needs to be. I think I might move the page back unless there are any objections. (Magnus: steal code, save headaches. See http://tavi.sourceforge.net especially, and also http://phpwiki.sourceforge.net )

Anyway, coding isn't much of a problem. Support within meta.wikipedia.com isn't necessary at the moment. I could help with the UseModWiki script. I might also finally implement MeatBall:PageDeletion too, which from the mailing list you apparently need.

On the more general issue, I'm not going to venture much. There are a lot of potential general arguments, but all ignore specific details that are salient. Specifically, at this time, you are faced with forking FOLDOC, specifically. Whether or not FOLDOC's existence is a threat to the very worth of Wikipedia I'll leave to philosophers. I'm not even capable of judging if Wikipedia is "worth" doing, as I never use encyclopaedia.

Nonetheless I will offer two things. Consider the extrapolation of your line of thinking. You are walking down the road where the only logical course of action is to accumulate all knowledge in your repository to the exclusion of others. I wouldn't support a project like that. It's not ethical, and it's hypocritical with the very goals of the GNU FDL.

Only in the Roman times has the worth of an encyclopaedia been to accumulate all knowledge, and I'd argue that was more harmful than beneficial. In modern times, an encyclopaedia is merely a synopsis of existing knowledge. It doesn't create new knowledge. An encyclopaedia exists to collate and organize information so a casual student may learn the basics of a topic, and an intermediate student may learn enough of a topic to search deeper through the literature of the given field.

If Wikipedia really wanted to be an Internet encyclopaedia "by the people", I think it would benefit from recognizing that context and taking advantage of the realities of the Internet. Wikipedia could serve as an organization of the information on the Internet as well, organizing other Internet resources (through URIs) within its structure. That would be really useful role given the "lost in a snowstorm" feeling many have.

But that's projection and futurism, not something I do. Right now, I think the chief benefit that Wikipedia has it that the participants have fun doing it. We could end up discussing whether or not it really meets the ideals of an encyclopaedia, and I'd point out that no encyclopaedia ever will be neutral or complete, but that's a scary postmodernism argument trying to burst out. Let's cap that oil fire right now, shall we? ;) -- SunirShah


It seems you're missing some context about the software change, why we're doing it, and why this wiki exists here at meta.wikipedia.com. Please don't make it sound as if we had not given it a lot of thought; we certainly have. You make a perfectly valid point, though, viz., that we could borrow code from other sources. For all I know, we already have done so.

You are walking down the road where the only logical course of action is to accumulate all knowledge in your repository to the exclusion of others. I wouldn't support a project like that.

Neither would I!  :-) But I don't see why having all knowledge in one repository--which is not exactly Wikipedia's goal, but no matter--would be to the exclusion of others. That it wouldn't be was precisely my point of my last contribution to this page! Wikipedia certainly is not an attempt to become the only free encyclopedia, but to become a comprehensive encyclopedia. Our being comprehensive does not entail our (absurdly) attempting to become the only free encyclopedia.

The rest is very interesting, but I don't have time to do it justice.  :-)

--Larry_Sanger


I agree with pretty much everything you say, Sunir, but the leadership has decided to avoid accountability and consensus, which leads inevitably to behavior which is autocratic and antithetical to the GFDL. This is one person for whom Wikipedia has ceased being enjoyable. --The Cunctator

I don't feel like getting into this again, but for Sunir's benefit, I want to point out that most Wikipedians do not share the opinion expressed here regarding Wikipedia's leadership. --STG


There now exists a second use case. PlanetMath. See http://www.nupedia.com/pipermail/wikipedia-l/2001-November/000791.html

So... what's the story, folks? Are we going to lose our momentum here? -- SunirShah


Sunir--it's not clear to me what you think we should be doing, or what momentum we had. Should we just ask people to write a script that will put up automatic links to FOLDOC (and now PlanetMath, it seems), when a Wikipedia topic is the same as a topic in FOLDOC and PlanetMath? I think that's a very interesting idea, and maybe indeed we should do it. Maybe you could make the proposal on Wikipedia-L, or do the coding yourself? --Larry_Sanger

I'm not going to get involved in the internal affairs of Wikipedia. As a non-community member, I am only willing to work at the periphery. If no one is willing to work on this, then it won't be done. I assumed someone was interested in trying this, but I no longer am sure that's true, so I'm not going to press it.

Meanwhile, the last couple weeks have seen some major improvements in wiki technology on MeatballWiki, like KeptPages controlled PageDeletion and file updating, and a RichSiteSummary mod_wiki extension, so I'm working on that. If someone wants to go further, then can contact me directly or ask on MeatballWiki where others may assist. -- SunirShah