[FM Discuss] wikimania '08 report

adam hyde adam at flossmanuals.net
Sun Jul 20 09:43:49 PDT 2008



hi,

[note: this is a draft, I will check URLs etc when I return to Amsterdam
and resend to the list. I thought it was of interest to sent his now in
any case]

So, Wikimania '08. It has been an interesting event. Held in the New
Great Library of Alexandria on the sea front of the beautiful
Mediterranean. The site is infront of the old Great Library which is now
under the sea.

I was here to represent FM and the good work we have all done, and learn
as much as I can from as many people as I can to further assist FMs
growth and development

I attended last years event in Taiwan so its interesting to compare. On
the whole I'd say this event has a taint more formality, due pretty much
to the venue. It is a prestigious housing for this event and while
impressive it does lessen the ebb and flow of casual meetings which was
a little easier last year in a more relaxed space.

Each days events were held across 5 parallel streams. There was no
attempt to harmonize streams, rather it seemed a little bit muddled with
no conscious emphasis on any specific themes. It felt a little bit like
a grab bag of ideas rather than a curated symposium. This feeling is
also enhanced given that each block is only 45 minutes which doesn't
give much time for discussion or in-depth investigation so I left many
presentations feeling that I got only a sampling of the promised topic. 

Even so, there has been some interesting stuff, so below is a short
report of what I found of interest.

First up, I missed the opening presentations as Brianna and I had to
prepare for a workshop later in the day, so the first thing I got to see
was Ruediger Glott and Rishab Ghosh (in the 3rd session) talking about
the forthcoming survey about motivations of Wikipedia contributors. I
have read many articles from Rishab Ghosh as he is one of the few clear
voices commenting on the motivations of free software developers and
free content contributors. You can find some of his work on the
excellent opensource.mit.edu resource for academic writings on open
source/open content.

Unfortunately however, the survey has not taken place yet so we did not
get the presentation of the results. This would have been OK, but
instead of an interesting drill-down into the issues at hand (motivation
etc) we got a brief overview of the survey itself, how it was formatted,
who should fill it out, and what it is trying to explore. The
presentation contained a disappointing lack of real substance.

Next up I attended the "Wikipedia Offline" presentation by Manual
Schneider. Manuel is working on the development of offline distributions
of Wikipedia on CD/DVD/USB drives etc. It was perhaps a little too dense
on the technical detail. I was also left with a nagging lack of
understanding for real world use for such a system. The work is quite
advanced, essentially focusing on optimising a usable (searchable)
offline snap shot of Wikipedia. However, there is no real work (as far
as I could see) into how to update the snapshots (eg writing to a flash
drive) or how to contribute and synchronise content written offline.
These two issues seemed to me to be the most interesting but were
entirely unattended to by the project.

Then we broke for lunch and afterwards Brianna and I lead a workshop on
using Inkscape with SVG and Wikimedia Commons. We had the joy of
competing with Jimmy Wales in his keynote. This fact was announced over
the intercom twice in Arabic and twice in English during the first 15
minutes of our workshop - but we weren't deterred! Brianna did the bulk
of the workshop, I demonstrated FM and introduced our series of Book
Sprints with special emphasis on the Inkscape Book Sprint and the
results, and then demo'ed Inkscape for tracing a bitmap. Brianna talked
about SVG, demo'ed the svgtranslate online service (more on this later)
and Wikimedia Commons. I was pleased Brianna and I had been slotted an
hour and a half (the only session to have such a generous allocation) so
we could get into some detail and discuss issues.

I think Brianna did a fabulous job and we maintained an audience of
about 40 people - most returning after the coffee break in the middle
which was a very good sign :) Congrats to Brianna, she did the vast
majority of the work and I think the audience was very impressed and
satisfied.

Incidentally - my room mate at the hotel (Carlos) is also part of the
organising committee for the next Wikimania (to be held in Argentina)
and I discussed the possibility of having more free software workshops
next year - it would be good to have workshops not only on free
softwares but on some of Wikimedia Foundation projects. Lets see if they
come back to us.

Next...the Wikimedia Foundation Board Panel. Seated like Egyptian
Pharaohs in a semi circle (remembering the legacy of the venue...) the
Board were actually pretty down to Earth. Its pretty great to see that
there can be an open panel discussion where the agenda is for the board
to update everyone on what has happened since the last Wikimania and
then open the mic for any questions on any topic. Remarkably egalitarian
in appearance although I am not close enough to the internal dynamics to
know how this is in reality. Still, its a good look.

Large on the discussion agenda was the question of a Volunteer Council.
This was explored in (sampled) detail in a panel the following day which
I also attended.

I left the first day feeling like I really didn't get much out of it
although a highlight was the workshop and meeting some friends I hadn't
seen for a while. However the content of the presentations seemed a
little superficial because of curation and the short length of each
session. 

I did have a great felafal (1 Egyptian pound, or about 15 euro cents).
and a few beers with Shunling Chen (FM Advisory Board) and a nice walk
along the stony storm break.

The next day it was back into it with an opening panel entitled
'Wikipedia Lectures'...the blurb read that it had something to do with
the organisation of lectures on all-things-wiki to be held in
irc...sounded good to me (I am involved in organising a 'sustainable
immobility' festival in Amsterdam in 2009 so I thought it might also be
interesting from this point of view). However, the topic advertised had
no relationship to the panel, although the panel was still very
interesting. 

4 old timers (Andrew Liu, Cat, Christopher Mathews + ?) presented their
point of views on the challenges facing Wikipedia due to the scale of
the project. One nice quote from a post (I think) from a wikipedian
named 'Durova' was put up on the beamer and seemed salient (the
following is paraphrased):
"At the beginning an organisation is based on relationships, as it grows
it develops policies until the point that these become too complex and
unsustainable, then the organisation reverts back to relationships"...

one of the panel suggested the last phrase should be "reverts to
politics"...an interesting point and it summed up a lot of the
frustration inherent in the points the panel was making.

The topic quickly narrowed itself down to the problems of becoming an
'Admin' in Wikipedia (there are 1500 Admins for Wikipedia English) and
how it is necessary to retain some kind of sane filtering process for
appointing admins but at the same time the sheer scale of the
organisation means the filtering process is now much more formal and not
as cosy as some would like. One panel member went as far as to say the
application process for Admins was emotionally traumatic and he would
not recommend anyone for an Admin because he didn't want them to go
through this process. Eeek! I don't see how these issues can be
harmonised, but once again (as per the board panel) it was refreshing to
see these issues discussed openly with no defensiveness apparent and a
short but good input from the audience. 

So, I was thinking - at last some interesting stuff! The rest of the day
built on this and I was very pleased in general with this day.

The next session was for Lightning Talks. These are a series of 5 minute
presentations on any topic. I got there early looking forward to seeing
something unexpected, and unexpectedly I was asked if I wanted to
present FM. I jumped at the chance and did a quick 5 minute presentation
on FM, an overview of the manuals, FM Farsi, and our Xchange (the tool
we use to transfer manuals between FM language sites) and Objavi (our
beta print on demand tool) tools. Afterwards I had a lot of good
feedback about FM so I was pleased to have had the opportunity. Of the
talks by other presenters I was most interested in points raised by an
Arabic speaker (I think he was Egyptian) about the need to grow the
Arabic Wikipedia and take pride in Arabic as the language (many Arabic
writers for Wikipedia prefer to contribute to the English site
apparently). It was very interesting discussion and great to see local
voices taking the floor - it turned into a passionate discussion that I
believe was followed up in another session.

After lunch I attended "Is the law an ally?" by French Wikipedia lawyer
Olivier Hugot. A very interesting presentation. His greatest piece of
advice is that if you receive a legal letter asking you to remove
content don't take it down until you have talked to a lawyer. This is
mainly because if you make efforts to remove the content this may
actually be seen as an 'admission' of culpability if the matter goes to
court. He also said the most common legal issues were misuse of
trademarks within Wikipedia, copyright infringement withing Wikipedia,
and defamatory comments within Wikipedia. 

Olivier believes the main threat to the WMF (he was talking mainly on
behalf of French law but was making a general point) would be a case
setting a precedent that considers the Wikimedia Foundation to be a
publisher. This is because publishers _can_ be held liable for the
content of their publications. He believes if this happened then it
would probably be a burden that could bring about the end of Wikipedia
as it would require unsustainable processes to filter content before it
made the public web. 

He likes to argue that the WMF is not a publisher but more of a channel
for other peoples content. The analogy he used was of talk radio or a
speaker in a venue. If a speaker says something defamatory then the
venue is not held liable. He believes this is the argument to make - to
classify Wikimedia Foundation projects as custodians of a channel, and
hence having no (or very little) responsibility or culpability for the
content posted.

In this way he tries to drive a wedge between 'channel' and
'publisher'. 

I asked him, but perhaps not in the most eloquent manner, at what point
might Wikipedia be considered a publisher. Specifically I was interested
in the new PDF creation tools wikipedia is experimenting with. The tools
are supplied by Wikipedia and enable users to create 'collections' of
pages that are then rendered as PDF. These tools, as far as I can tell
through talking with Eric Moeller (involved with developing these
tools),  is to tie Wikipedia into print on demand services. Given that
ebook readers are on the rise (some read PDF) and PDF is the source
format for print on demand - does this tip Wikipedia over this fine line
into the role of a publisher?

However I don't think he understood my question so I can't report any
interesting conclusion on this. ugh. Maybe Shunling would like to
comment on this?

Next...Brion Vibber - lead developer for MediaWiki (the software that
drives all Wikimedia Foundation projects). His presentation was weirdly
similar in feel to a Steve Jobs MacWorld presentation (although
possibly a little geekier). Brion is something of a star in this arena.
First up some very interesting stats:
* across all WMF projects there are 10,000,000,000 page requests a month
* there are 50,000 http object requests per second
* hardware for hosting costs 1.5 million (USD) 
* 25,000 USD is spent on bandwidth per month
* hosting (housing) costs $10,000 a month
* there are 4 paid programmers/system administrators and 3 volunteers (a
herculean ratio)
* all servers are standard x86 (64 bit) servers running Ubuntu

The presentation was, once again, rather upfront. His summary is that
Wikimedia Foundation is now out of the 'panic mode' and is now a
sustainable technical enterprise. He outlined some of the most recent
achievements as being:
* single user login (you can create one account which can log you in
across all WMF projects).
* an experimental mobile version (you can see it at
http://mobile.wikipedia.org)
* a new localisation service for WMF projects
(http://translatewiki.net). I believe Wikimedia projects use 'gettext'
for localisation which is a unix tool that creates a specific kind of
format for translation strings (sometimes called the 'gettext
standard'). FM uses the same syntax and I was cornered by one of people
involved in this translatewiki project to see if FM would use the
service...no need as we have our own very nice tools (see the
localisation chapters in the FM manual)


Things that are not on the cards for MediaWiki include:
* no WYSIWYG editor. this is largely due to the fact that WMF syntax
(mark up) is also used for formatting templates and it is hard to break
the two issues apart.
* no p2p wikipedia (an ongoing, old, issue)

In the pipeline:
* the PDF generation is still in the works but in beta mode. Largely due
to syntax rendering as far as I can see. I believe the developers are
called 'PediaPress' and they have some information on their site
although I haven't checked it yet.
* bulk upload and simplified meta data entry for Wikimedia Commons
* possibility of embedded ogg theora in wikipedia articles. They are
working closely with Metavid (http://www.metavid.org) on this.

Next was the Volunteer Council meeting. This is largely about the
internal workings of Wikipedia. The Wikimedia Foundation has a board,
which is soon to have two chapter members (a 'chapter' is a language
project) and two community members. It also has some other elected
members including a chair and the 'honorary' position for Jimmy Wales.
The board promote themselves as outside of the daily operations of
Wikimedia Foundation projects and say they actively try to 'get out of
the way' of Wikipedia et al. I can't comment on whether this is actually
the case. However there seems to be murmurings within the
'community' (the nature of this in itself is hard to define) towards the
establishment of some kind of Volunteer Council (VC) that will represent
concerns to the programmers (who are mostly paid staff) and the board
etc.

What the mandate is for the VC, what it should do, how it is
appointed/elected, and how it operates is all in the air. So this
meeting was really a realspace catch up and attempt to synthesize
discussions that have occurred online.

Brianna may wish to comment on this further. I found that the discussion
was interesting but it was pretty much impossible for it to get very
far.

Lastly I attended a split presentation about 'the wikisation of images'
by Nikola Smolenski and 'metavid'  by Michael Dale. Both were technical
presentations. Nikola spoke about the need to develop a wiki syntax for
image creation. The idea is that images could be augmented/scaled etc
which as defined by the wiki mark up. In addition he spoke about the
ideas Brianna has been discussing in our workshop about how SVG is the
most 'wiki-like' image format in that it is a editable text file and so
its easy to manipulate programmatically.

Nikola developed his svgtranslate script which is online and it enables
you to enter a url for a svg, it then parses the svg for text components
(these are stored as plain text in the file) and then creates empty form
fields for each text block. You then can enter new texts and the script
will replace the old text with the new. The final result being an SVG
with the new text. Great for translations etc. Brianna had previously
demonstrated this during our workshop but it was great to see the guy
that made it present the rationale etc.

Nikola had also worked on scripts that will take regional maps and add
highlighting to specific areas according to the mark up used. For
example you can produce a map of europe with any country highlighted,
all created in the fly and all done via SVG templates and wiki mark up.

It was a great presentation and really opened my mind to the
possibilities for SVG. 

Next was Michael and Metavid. Metavid started as a US Congressional
video archive with very sophisticated indexing and searching.

However metavid is now a content agnostic video 'cms' (wiki) of sorts.
It is based on Mediawiki and Semantic Mediawiki but it has really taken
it to a whole new level.

The feature set is extensive but some highlights include:
* inbrowser video editing
* potential for integrating speech -> text transcripts
* inbrowser subtitle (transcript) editing
* huge list of supported codecs 
* smart plugin detection and manipulation
* time based url syntax 
* ability to embed 'snippets' using their url syntax and embed api

and more...

A hugely impressive project. My only, and minor criticism, is
that it lacks good design. Infact it falls into the category of bad
interface design. This is really a shame and I hope they address it as
Metavid is a fantastically extensive and sophisticated approach to video
hosting.

In the evening I met with Samuel Klein from OLPC about the documentation
of OLPC and Sugar in FM. It was an interesting discussion, more about
this in another email. Also I received an email from Lotte saying the
book Brianna wrote during the sprint (see the new Wikimedia Commons
manual linked from the front page of the English FM) arrived and looks
beautiful. This is the first book made using our Objavi automated
FM->print pdf generator. Cool! Can't wait to see it :))))) With luck one
might arrive before we leave Wikimania but its unlikely...

So! Day two, was done. Much more interesting than day one but still
needed more time in each of the sessions.

Next day (day 3)...Mako (FSF) started with a presentation about free
network culture. A short introduction into the issue of network services
and what 'free' means in this context. What does 'free' mean for network
services like Google, Facebook etc...do we talk about privacy, free
code, free information, transparency? The GPL v3 does not address this
critical issue although there is the AGPL which is considered by some to
go someway towards addressing the issue but not far enough. It was a
great presentation but no time for questions.

Then there was a (hohum) Creative Commons panel. ra ra ...CC is a dull
topic these days if not given some critique. The panel focused just on
what CC is, and Mako introduced how CC applies to Wikipedia. In short,
CC doesn't apply as Wikipedia is covered by the FDL (which is a License
that masquerades as 'free' but is infact a jail for content). Its an old
story but essentially Wikipedia want out of the FDL so they are trying
to get the FSF to make the new draft of the FDL compatible with the
CC-BY-SA license. This would mean Wikipedia could slip out of the FDL
and into CC. Questions remain - should the nature of a license be
influenced by the needs of one user of that license? NO is my answer.
They have done this kind of stupid thing before with the FDL. The
rationale for the license includes a paragraph about how the license was
originally written to protect the business interests of small publishers
that wanted to publish manuals on free software. It is silly, if you ask
me, to write a free license to such a specific context. It looks like
the FSF is about to do the same thing again by changing the FDL to suit
the interests of the Wikimedia Foundation.

I believe the FSF should just think of the license and its needs. What
constitutes a free license in this context? In my opinion a really free
license for documentation (and I actually don't think there should be a
special license at all for documentation) is the GPL. It does the
job..whats the problem? By having a non-free license paraded as free in
the FSF license stable just weakens the trust in the FSF. If something
isn't free, and the FDL does not preserve freedom, then two things need
to be done: first - don't call it free, second - dump it.

What the FDL needs is to be written so that it is scheduled for being
outmoded and give anyone that uses it the opportunity to stay with the
old outmoded license, or move to the GPL. But...no...no time for
questions...time taken up by someone asking if they should use a
CC-BY-SA-NC or CC-BY-SA license...note to next programming committee :
hold intro courses on all WMF projects and free licenses so that panels
don't get stuck on the basics.

Next....Brianna on Wikimedia Commons. You may not know Wikimedia
Commons. If you don't then read Briannas manual
(http://www.flossmanuals.net/wikimediacommons) Essentially Wikimedia
Commons is a media library for Wikimedia projects. It has a special
focus on images but also extends to audio and video.

Some interesting notes:
* 5000 files are uploaded to Wikimedia Commons daily.
* 3 million files are in WC as of the day before the conference began
* the contributions are growing faster than new articles in the english
wikipedia 
* the single user login is expected to accelerate this uptake even more


Brianna gave an overview of WC and where it is going. The point I noted
most of all was that WC is the only WMF project that works with all
other WMF projects. This is because many of the WC images get used in
the WMF projects. 

It seemed to be that next to Wikipedia english (arguably the flagship
project) Wikimedia Commons is pretty much the next most important
project on the table. Although this may or may not be the case
(interested in Briannas thoughts on this) it certainly seems that its
importance is not reflected in the attention accorded to it by the WMF
or its programmers. As I understand it, from the bug list presented by
Brianna, the project has some severe usability issues that hinder its
uptake and usefulness. Most notable are:
* you cant rename/move images
* there isn't a workflow management system for classifying/clearing
images
* the upload process is just too hard to use
* there needs to be clean batch uploading
* it isn't really usable in many languages other than english

Also, it seems the default set-up for all new WMF projects enables users
to upload images etc to the 'local' project and hence many images are
being uploaded to each projects repository instead of to the shared WC
repository. It would seem reasonable to require new projects to have
this 'local uploads' feature turned off by default so that all projects
upload to WC and share images etc.

They seem like pretty critical issues to me. On the up side is that
despite all of this, WC is still being used by a tremendous number of
people.

Ok...moving on...WikiTravel presentation...WikiTravel are a wiki
(mediawiki) for travel guides that dovetails into a print on demand
process. Interesting interesting. Very similar themes raised to those
here at FM...

First...WT pay editors to write and edit documents. As I understand it
anyone can contribute but the editors get paid to maintain the bulk of
the content and ensure its clean and ready to print. Paying people for
content goes against the flow in the WMF world (at the moment at least,
although there was some mention of grants from the WMF for developing
people etc but I'm not sure that is the same thing). 

The received wisdom in this is that paying people decreases motivation
and increases discontent within those that don't get paid. WT believe
there is no problem paying people (and I agree). WT believe the role of
the paid contributors is clear and specific. It is WTs belief that this
is enough to justify paying them and not others. I don't think I quite
agree with this. My feeling is that it is a matter of organisational
culture. If its clear from the beginning some people might be paid (and
it could be anyone) for some task, but volunteering is also encouraged,
then I believe this can work well. We have been doing this within FM and
so far I haven't heard any grumbles...(speak up if there are!)

Next WT investigated what do 'real travel writers' think of contributing
to WT. Answer : they hate it. Well, that's upfront I guess. According to
WT professional travel writers hate using a wiki. They think the mark up
is too hard and would prefer WYSIWYG editors or, preferably, to write
material offline in word processors. WYSIWYG is too difficult to
implement in MediaWiki (as per Brion Vibbers comments) so that's not
going to be solved early. Also, travel writers apparently feel a loss of
ownership of the content (ie. their perceived authorship is diminished
if the document is written collaboratively). 

So WT don't worry about the professionals but choose to work with 'non
professionals' that are comfortable in wiki-space. Interestingly,
professional documentation writers do use FM - how do you feel about
this issue Anne/Elisa/Cedric (and others)?

Next...how does the commercial book market react to print on demand?
Answer : not very well. According to WT, book sellers do not recognize
'wiki output' as 'books'. I am not sure I agree. There are some computer
book publishers (PAKT) that write (closed source) manuals and publish
via print on demand. Book sellers buy them...seems to be no problem. I
also asked one book seller in Amsterdam about the fact FM produces
content in a wiki, he had no problem with it and would consider buying
books produced this way. 

One issue however is that of ISBN. These are the numbers that identify
each book. An ISBN costs $225 (USD). If you release a new edition you
need a new ISBN. ISBN is used by all book buyers and distributors for
inventory control and ordering.

The big hang-up here is that content produced in a wiki and delivered
via print on demand doesn't really hold with the idea of 'editions'. If
you are editing in a wiki and updating the print source constantly, then
what on earth is an 'edition'?

So...is there an alternative? Well...ISSN exists. Which is ISBN but for
serial publications like magazines. So is this ok to use? It seems,
according to WT, that book sellers think magazines are magazines and
books are books, and books cannot share the same identification system
that magazines have. That would be just too confusing for the poor book
buyer.

So..it could work, but might have some resistance by some book buyers...

There is also the Amazon Standard ID Number (ASIN). But this is only for
distribution via Amazon partners, hence if you use this you will never
get your book in Barnes and Noble (etc).

WT advocates a free ISBN/ISSN/ASIN alternative. Sounds good.

I asked them in question time about their PDF rendering process. They
need to get from MediaWiki to Print and PDF is the intermediary format.
Ok...is it open source? No. Ok...is there plans to make it open source?
Sort of. (right). I don't think this is a good look.

Then later the PediaPress chap stood up and said they provide the tech
for WT and it is open source. I asked him afterwards for a demo and the
source url and he delivered both. Goodo (->http://code.pediapress.com) 

Ours is better ;) (yay Objavi!)

Ok...last up - end of event pat on the back to everyone (literally), group
party and final goodbyes.

So...my impression overall..it was a good event. I enjoy Wikimania
mainly for the people. You learn more talking to the person sitting next
to you at lunch than from most presentations. With regard to FM - if I
can at all talk on behalf of others I met then I would say that people
are impressed by FM and what we have done. I met many that love the what
we do and some that have used our manuals and benefited from it :) I
feel very proud representing the project here and I learned a great deal
which will I believe will help us improve what we do.

As for Wikimania improvements.....I think that WM could certainly be
improved. Mainly I would say :
* longer sessions
* workshops on WMF projects and licensing
* clearer classification of sessions (you would have thought that the
WMF of all ppl might have got this right...)
* ensuring all presentations were about the subjects advertised
* a better communal hang out space
* power plugs for all

Ok..those are the bugs. On the whole, a very worthwhile and enjoyable
event. As last year, I enjoyed the openness of everyone and the
generosity of spirit. Its a pity more projects don't come that are
interested in Wikis in general as I found that while the event is
obviously WMF-centric I found everyone willing to talk about other
projects outside this scope and free culture in general.

On a note about the people...the event was great for meeting people as
it was last year. However I have to say, the local Egyptian people are
incredible. Very open and generous and very very kind.

all for now :)

adam

























































More information about the Discuss mailing list