Hello, I regularly read and contribute to Wikipedia under the name Iammaxus.
Over the last few weeks, I have been mulling over an idea on how to
significantly improve the mediawiki project. I posted the following on my
user page before finding and reading this mailing list only to find out that
this idea has been discussed here
(http://mail.wikipedia.org/pipermail/wikitech-l/2003-December/007185.html),
albeit not with the scale that I'm imagining Because im not sure where to
put this in the metawiki, and im not sure if anyone would care if I did,
here is an excerpt from a conversation i had with a friend of mine
knowledgeable in computer science on 12/24/03 (note that I wasn't seriously
asking him to do this, just more of bouncing an idea of off him):
I Am MAXUS (2:30:14 PM): yo
Chin Jut: hi
I Am MAXUS: code a meta data system for wikipedia
Chin Jut: what's that mean?
I Am MAXUS: i can't even begin to imagine the possibilities
I Am MAXUS: well first off
I Am MAXUS: organize topics
I Am MAXUS: in a tree system
I Am MAXUS: this would be part of the meta data system
I Am MAXUS: well wiat
I Am MAXUS: lemme start from the beginning Chin Jut: ok
I Am MAXUS: the overall point is to allow more machine generated info, stuff
that shouldnt be manually made like it is now, or even with one time use
scripts
I Am MAXUS: such as lists of articles
I Am MAXUS: tables of dates
Chin Jut: alright
Chin Jut: "List of famous bears", that sorta thing
I Am MAXUS: but much greater than taht
Chin Jut: alright
I Am MAXUS: so u could just request a list of a certain sub tree
I Am MAXUS: etc
I Am MAXUS: then it would have more specific meta data
I Am MAXUS: such as meta data about books including the author and such
I Am MAXUS: so that this would automatically be put into an article about
the book
Chin Jut: ok
I Am MAXUS: so:
I Am MAXUS: ?
I Am MAXUS: go
I Am MAXUS: ]and do it
Chin Jut: well, gee
Chin Jut: that's a large project
Chin Jut: I'm not even sure what to begin with
Chin Jut: I mean, what features need to be available?
Chin Jut: (I'm not gonna be able to do this, why am I talking?)
Chin Jut: what, concretely, needs to be done? I Am MAXUS: good question
I Am MAXUS: well first of all, learn xml and shit
I Am MAXUS: cause thats how all this junk is done
I Am MAXUS: or somethintg
I Am MAXUS: lol
Chin Jut: But Tim Sweeney speaks disparagingly of XML...
I Am MAXUS: does he? Chin Jut: yeah
I Am MAXUS: what does he say is bad about it?
Chin Jut: lemme see if I can find it
Chin Jut: "Does anyone else see XML as an overcomplicated solution the
meager problem of serializing data in and out of text files?"
Chin Jut: Philip Wadler (one of the main guys behind Haskell) also bashes
XML: "So the essence of XML is this: the problem it solves is not hard, and
it does not solve the problem well."
Chin Jut: All the same, yeah, I'll learn XML
I Am MAXUS: lol
I Am MAXUS: well screw those guys
I Am MAXUS: cause everyone is using it
Chin Jut: yeah
Chin Jut: Phil Wadler goes on to say
I Am MAXUS: waht does he mean "serializing data..."?
Chin Jut: It's worth studying XML just because it became popular while
better things did not
Chin Jut: Serializing data means writing it and reading it from files
Chin Jut: you take a complicated structure, like a tree
Chin Jut: and turn it into some linear sequence of bytes
Chin Jut: hence, you've turned it into a series... you've serialized it
I Am MAXUS: right
I Am MAXUS: anyway i dont know if u can usel xml
I Am MAXUS: because of the database based nature of it
I Am MAXUS: but something similar
I Am MAXUS: anyway
I Am MAXUS: i dont know about searching, and thats the main function of this
system
I Am MAXUS: so u have to figure out how to do that
I Am MAXUS: so each article has info attached to it
I Am MAXUS: ideally, the types of info could be specified by ppl in a
realtively plain language way
I Am MAXUS: so taht ppl could specify more types of meatadata for certain
types of files
Chin Jut: what types of metadata would people specify?
Chin Jut: Like "this article is about bears"?
Chin Jut: Shouldn't that be auto-discovered by computers?
I Am MAXUS: thats beyond the scope of this
I Am MAXUS: that requires all sorts of human communication stuff
Chin Jut: ok
Chin Jut: so then what is this, exactly?
I Am MAXUS: i told you!
I Am MAXUS: for example, lets take the organism pages
Chin Jut: ok
I Am MAXUS: they all have the classification on the side
I Am MAXUS: well instead, each page would have a is a member of this higher
group
I Am MAXUS: which wikipedia would look at
Chin Jut: I see
I Am MAXUS: and see what thats a memeber of
I Am MAXUS: and dynamically figure out the whole classifcation
I Am MAXUS: this is not such a useful example because its much more static
I Am MAXUS: but take the date pages, those are useful mommas to meta-fy
Chin Jut: ah
I Am MAXUS: so if a page is an event
Chin Jut: so people would have to say in the event
I Am MAXUS: it would include date info
I Am MAXUS: and type of event info
Chin Jut: "Date info: July 4, 1776"
I Am MAXUS: right
Chin Jut: and then the date page would say "Search for all pages with date
info: July 4, 1776"
I Am MAXUS: so it coud be added to the "type of info in history" page
I Am MAXUS: that too
Chin Jut: "typo of info in history" page?
I Am MAXUS: well like the music in history page
I Am MAXUS: etxc
I Am MAXUS: etc
Chin Jut: ah
I Am MAXUS: so if u had a page that was under the music subtree
I Am MAXUS: and then under the band subtree
I Am MAXUS: and u had dates of existence of that band
I Am MAXUS: and then there could be an important concert subtree, etc
Chin Jut: well, to be technical, I don't think these are trees, I think
they're DAGs
Chin Jut: but it doesn't matter
I Am MAXUS: dag?
Chin Jut: directed acyclic graph
Chin Jut: in a tree, a node has only one parent
Chin Jut: (at most)
I Am MAXUS: yeah good point
Chin Jut: ok, it's sorta interesting. I have no idea how to do it
efficiently, though. But I think I might actually work on it
I Am MAXUS: lol
I Am MAXUS: i want to copy this conversation into the metawikipedia.org
somewhere
I Am MAXUS: maybe in the todo for vers 4 or 5
Chin Jut: I want to eat breakfast, because I have yet to do so
Chin Jut: bbl
I Am MAXUS: bye
Chin Jut signed off at 2:48:15 PM.
Those are our AOL Instant Messenger screen names. His wikipedia name is
Chinju
Thanks for reading this, I hope you consider this because I believe this
idea has the potential to revolutionize not just the wikipedia, but all
sorts of projects.
I think we should do this ASAP, because the issues are only going
to get harder. But the main issue, Timwi, is social -- there are
certainly cases where two different people have the same username
but in different languages.
Timwi wrote:
>
> Thomas R. Koll wrote:
> >On Mon, Dec 29, 2003 at 11:04:00PM +0100, Erik Moeller wrote:
> >
> >>Stuff like single sign on
> >
> >That is not possible anymore. It would be total chaos.
>
> I'm pretty confused at this assessment of yours.
>
> All it would take to achieve this would be to:
>
> - Add a field to the 'cur' and 'old' tables (or perhaps even just the
> 'cur' table) to specify what language an article is in.
> - Move all the articles from all languages into one database.
> - Adapt code accordingly.
>
> Should be doable (given adequate manpower).
>
> Timwi
>
>
> _______________________________________________
> Wikitech-l mailing list
> Wikitech-l(a)Wikipedia.org
> http://wikipedia.mormo.org/mailman/listinfo/wikitech-l
>
I don't suppose that there is a history graph of the donations to Wikimedia.
I was currious
how much of the US$31k was raised as a result of Jimmy Wales' December 28
letter and the
associated Slashdot coverage. I was also wondering if Wikimedia's
accounting ledegers are
open to the public. If not, why (this isn't a flame, I'm sure that there is
a good enough
reason)?
Thanks for the great site. In my opinion Wikimedia deserves at least
another US$30k in 2004.
--adam
--
"Any society that would give up a little liberty to
gain a little security deserved neither and
will lose both." --Benjamin Franklin
"Necessity is the plea for every infringement of
human freedom. It is the argument of tyrants; it
is the creed of slaves." - William Pitt
--
+++ GMX - die erste Adresse für Mail, Message, More +++
Neu: Preissenkung für MMS und FreeMMS! http://www.gmx.net
For the past few weeks, I and some others have been in the process of
setting up a new MediaWiki-based wiki site devoted to the "Star Trek"
television series, called Memory Alpha. We have had few problems after
the initial pains of actual installation; however, in trying to adapt
the MW framework for our site, we've been running into a few snags.
The main problem from my perspective is documentation for users of the
wiki software in sites which are not straight-up mirrors of Wikipedia.
We've been doing a rather haphazard job of copying over the various
bits of documentation for using the site -- stuff like how to write an
article, how to edit a page, the policies and guidelines, and other
material.
Our problem is that so many of these pages will definitely be useful,
and although they'll most likely be adapted and changed as time goes on
and we set up our own policies, I think it would still be useful if we
could have something that we could start with.
Currently, MediaWiki uses the website's name to create a separate
namespace for documentation and other "meta" type pages. (I'm not
referring to the Meta-Wikipedia, but rather the "Wikipedia:" pages on
Wikipedia itself.) And I realize that this is probably a workable
system...
But, considering that the MediaWiki software is made publicly available
for download and for establishing other websites, I wonder if it might
be useful to have some kind of help "module" -- that is, a collected
copy of the documentation pages that can be easily copied and set up on
other sites. Possibly, this could also add a new "Help:" namespace
which would help distinguish those pages for the general users. But
that last isn't really necessary as far as the content goes, although
it would make linking between the various help pages simpler if the
namespace wasn't changed on each site.
I'm not a programmer myself, although I've got a basic grasp of coding
and I'm very slowly learning PHP. I could probably try to help if
someone else wanted to try to implement this or something similar, but
I thought that I would propose this idea to the group first, seeing as
how many of you are a lot more familiar with the MW's inner workings.
Thanks,
Dan Carlson
Administrator, Memory Alpha
http://memoryalpha.st-minutiae.com/
I did some googling tonight and found some more information about ways to
do some magic distributed caching.
An example for a proprietary product doing this is Cisco
DistributedDirector (19000$). Anybody?
Ok, so here is the open source alternative:
* Super Sparrow http://www.supersparrow.org/
Open Source, linux, and tested. It's running for example vergenet, you can
see it in action at http://www.vergenet.net/vergenet/.
In combination with Linux Virtual Server& Heartbeat plus distributed squid
'mirrors' this looks like a nice way for future growth.
IMO this is nothing for the immediate future, but good to keep in mind and
start playing with. Maybe it would also be possible to ask Horms (Simon
Horman, http://www.vergenet.net/~horms/) for advice, he definetely is an
expert in this field.
Have a nice new year!
Gabriel Wicke
Jake Nelson wrote:
> The thought's crossed my mind a few times as well- they've got very good
> connectivity, and IIRC, it's about $5k for a colo rack slot, which isn't
> unreasonable for being outside of national jurisdictions.
It might not be unreasonable if we were doing anything illegal or
controversial. But we aren't. If anything, I'm going to be
struggling with my own moral reservations about government funding
rather than worrying about governments shutting us down. We're just
the sort of feel-good project that politicians love to take credit
for.
Because of our license, copies of Wikipedia's database are already on
probably thousands of machines all over the world. The servers are
based in the U.S., where the First Amendment as well as the general
culture give us broad protection and political support.
I think Havenco is neat, but for us, it's really not desirable.
--Jimbo
Hi,
i've googled around a bit about possible 'global load balancer' setups.
A proprietary thing that does this is Cisco DistributedDirector. It's
not cheap- i've seen a price of 19000$. Nice if Cisco donated one...
However, there seems to be a good Open Source alternative at
http://www.supersparrow.org/.
I'm into party preparations, have fun tonight!
Gabriel
Nikos-Optim wrote:
> I think wikipedia should look the same, have the same
> policies and customs, and the same spirit, no matter
> the language.
In terms of very high level policies, I agree with you. But there are
lots of little things that are dependent on culture. Let me give a
couple of examples.
1. I am told that in some languages, what is known as an encyclopedia
tends to include more 'mere definitions' that in English would
properly belong in a dictionary. So, if the French wikipedia chooses
to draw the line differently than we do about that sort of thing, I
don't see a problem with that.
2. It has been reported to me that on the Japanese wikipedia, there
is a stronger tendency to discuss things first on the talk page,
reaching a consensus, and then to update the article page. That's
consistent with Japanese culture. In en, of course, there's more of a
custom of boldly updating pages and while simultaneously arguing about
it on the talk pages. Either custom is valid, if it suits the people
who are doing the work.
--Jimbo
> > Jason's got Ursula back up, and our new machine is also installed. I'm
> > copying files over so it can take over pliny's web work and let Ursula
> > do just the db.
>
> Out of curiosity, why did Ursula go down? If the cause is unknown, could
> there be an issue with our database that might cause such crashes?
If machines keep dying, there could also be weird stuff happening with power
supply.
z.
On Wed, Dec 31, 2003 at 03:18:28AM +0000, Timwi wrote:
>
> Ray Saintonge wrote:
>
> >Nikos-Optim wrote:
> >
> >>I think wikipedia should look the same, have the same
> >>policies and customs, and the same spirit, no matter
> >>the language.
> >
> >That sounds like a very effective policy for driving people away.
>
> That sounds like a pretty bold statement without any real back-up ;-)
The Esperanto has uppercase-letters for family-names, only one example.
Asian, Hebrew and Arabic Wikipedias need a different look for the Interface
because they don't write left->right like we do.
> It would certainly not drive *me* away -- quite to the contrary, it
> would encourage me to contribute more on non-English Wikipedias.
I doubt you really have enough time to contribute to more than one or
two WPs.
--
http://www.tomk32.de - just a geek trying to change the world
-.- http://de.wikipedia.org/wiki/Benutzer:TomK32
/|> http://tomk32.bookcrossing.com
/ \ http://tinyurl.com/u6de