Redesigning the WMF from scratch

[Vigilant]

In another thread, I missed an excellent question from Carcharoth:

Bonus question: if you were designing the WMF from scratch, how many people would you employ and how would you organise it?

Since we're at the organizational crossroads and I'm the devil waiting to barter for souls at midnight, let's take a stab at it.

As the basis for the expounded conversation, I would posit that the WMF needs to do three things well; everything else is ancillary:

1) Keep the servers running
2) Present a professional and sympathetic face to the world to keep the money flowing
3) Prevent wikipedia from turning into a dead end property (AoL, MySpace, etc, etc, etc)

Thoughts?

[Zoloft]

Stuff that needs fixing, spurting arterial blood:

• Wikipedia is stuck with rules and pillars that were designed for rapid expansion and adoption, not sustainment and refinement of content.
• Wikipedia culture cold-shoulders new members. Most leave.
• Experts have a difficult time surviving in the Wikipedia environment, much less thriving.
• MediaWiki is a shambling monster of non-portable markup and creaky code.

[Malleus]

Stuff that needs fixing, spurting arterial blood:

• Wikipedia is stuck with rules and pillars that were designed for rapid expansion and adoption, not sustainment and refinement of content.
• Wikipedia culture cold-shoulders new members. Most leave.
• Experts have a difficult time surviving in the Wikipedia environment, much less thriving.
• MediaWiki is a shambling monster of non-portable markup and creaky code.

Everything springs from #1 I think, and its subsequent ossification. The MediaWiki software could easily have been updated by a few competent programmers, and still could be, but the WMF didn't recruit competent programmers, as evidenced by the Visual Editor debacle.

So as to a redesign of the WMF, keep Jimbo Wales as far away from it as possible.

[EricBarbour]

Put all of it together and make a blog post.

[Malleus]

The WMF is clearly an incompetent software development organisation, but its Wikipedia flagship has at least as many problems. So maybe the question ought to be "If you were going to redesign Wikipedia ...". It's noticeable that Zoloft didn't distinguish between WM and WP in his posting for instance.

[Peter Damian]

Stuff that needs fixing, spurting arterial blood:

• Wikipedia is stuck with rules and pillars that were designed for rapid expansion and adoption, not sustainment and refinement of content.
• Wikipedia culture cold-shoulders new members. Most leave.
• Experts have a difficult time surviving in the Wikipedia environment, much less thriving.
• MediaWiki is a shambling monster of non-portable markup and creaky code.

You have my vote next Committee elections. And I actually can vote this time.

[When pigs fly]

In another thread, I missed an excellent question from Carcharoth:

Bonus question: if you were designing the WMF from scratch, how many people would you employ and how would you organise it?

Since we're at the organizational crossroads and I'm the devil waiting to barter for souls at midnight, let's take a stab at it.

I'll steal a joke from the Vietam war era. First put all of the good editors in a boat. Blow everyone else to smithereens, then sink the boat.

WMF would be better off ceding control to the University of Pittsburgh who ran the excellent Telefact program which sadly died due to funding cuts. They would do FAR more with less and put the funding revenues to better use. Carnegie Mellon would also be an excellent candidate.

Why are people surprised when you let amateurs run something, you will get amateur results?

[arthur]

WMF would be better off ceding control to the University of Pittsburgh who ran the excellent Telefact program which sadly died due to funding cuts. They would do FAR more with less and put the funding revenues to better use. Carnegie Mellon would also be an excellent candidate.

Or Rice University:

Connexions is an environment for collaboratively developing, freely sharing, and rapidly publishing scholarly content on the Web. Our Content Commons contains educational materials for everyone — from children to college students to professionals — organized in small modules that are easily connected into larger collections or courses. All content is free to use and reuse under the Creative Commons "attribution" license.

[When pigs fly]

WMF would be better off ceding control to the University of Pittsburgh who ran the excellent Telefact program which sadly died due to funding cuts. They would do FAR more with less and put the funding revenues to better use. Carnegie Mellon would also be an excellent candidate.

Or Rice University:

Connexions is an environment for collaboratively developing, freely sharing, and rapidly publishing scholarly content on the Web. Our Content Commons contains educational materials for everyone — from children to college students to professionals — organized in small modules that are easily connected into larger collections or courses. All content is free to use and reuse under the Creative Commons "attribution" license.

CMU handles the software, Rice content control and a third unknown entity handles policing.

I assume the BOD could with a simple vote make this happen. Lord knows all three of these highly respected institutions would fall over themselves to take control of such a highly regarded project with an obscene level of funding.

So let's pretend this would be Utopia, and all of these institutions would want to play ball. Why wouldn't the current board do this? Do they have a relationship or interest with WMF employees living high off the donor's tit?

[Neotarf]

WMF would be better off ceding control to the University of Pittsburgh who ran the excellent Telefact program which sadly died due to funding cuts. They would do FAR more with less and put the funding revenues to better use.

From Pitt eliminates TeleFact funding stream

SGB President Gordon Louderback said that the Board decided that TeleFact should no longer qualify as a formula group because technological advances led to a decline in the usage of the service. He also said that a majority of its funding went toward paying employees.

“There are only a handful of [formula groups], and we as a board felt that TeleFact was no longer one of those organizations as it was now rivaled by the availability of the Internet on computers and smartphones,” he said.

Bonner reiterated Louderback’s claims, saying the decrease in usage necessitated a change in funding strategy.

He also said that according to data TeleFact collected and presented, annual calls to TeleFact have decreased by approximately 97% since the 2008-2009 academic year, when annual calls totaled approximately 100,000.

So they paid a bunch of people to read the internet over the phone. And now it's faster for people to google it themselves than to call up someone for the answer.

And the budget?

TeleFact received 3.4 percent of the Student Activities Fund, a $2.3 million fund collected from the Student Activity Fee paid by each undergraduate student.

[When pigs fly]

Telefact's demise is definitely due to the Internet, but Telefact was not created to provide answers but rather train researchers, especially Librarians. Those kids at Pitt were trained far beyond than the average slob using Google.

[Neotarf]

Telefact's demise is definitely due to the Internet, but Telefact was not created to provide answers but rather train researchers, especially Librarians. Those kids at Pitt were trained far beyond than the average slob using Google.

It's amazing what can be done with the Reader's Guide and a few fiche readers, isn't it. But this looks a lot like your proposal to pay Philippine school girls to make edits. And tsk, you didn't even get consensus before declaring it to be a GGTF group proposal.

[Zoloft]

Just to let the participants in this topic know, I will tomorrow mercilessly weed out off-topic posts. I appreciate people who don't leave me a lot of work.

If follows logically that the opposite also applies.

[EricBarbour]

It's amazing what can be done with the Reader's Guide and a few fiche readers, isn't it.

If more "Wikipedians" did that, rather than lazily Googling everything, Wikipedia would be quite a lot better than it is.

[Neotarf]

It's amazing what can be done with the Reader's Guide and a few fiche readers, isn't it.

If more "Wikipedians" did that, rather than lazily Googling everything, Wikipedia would be quite a lot better than it is.

This is the whole point of GLAM and the editathons. Google has the advantage of giving sources that can be checked easily, but when these institutions open their collections for groups to use for editing, this can add some depth to an article, as well as establish notability for more specialized subjects.

Stuff that needs fixing, spurting arterial blood:

• Wikipedia is stuck with rules and pillars that were designed for rapid expansion and adoption, not sustainment and refinement of content.
• Wikipedia culture cold-shoulders new members. Most leave.
• Experts have a difficult time surviving in the Wikipedia environment, much less thriving.
• MediaWiki is a shambling monster of non-portable markup and creaky code.

The GLAM model solves several of these problems; it accommodates academics who would otherwise be driven off by the caustic editing environment, it welcomes newbies, and provides them with coffee and experienced users to get them over the initial small hurdles. And if you haven't checked the visual editor lately, take another look. If you experienced it in its most horrible alpha form, chances are you will never warm up to it, but newbies take to it well, even though it still has some rough edges. Rules and pillars you don't need for a group that only spends a few hours editing. You only need a few policies like RS and NPOV. The drawback of this kind of editing is of course that everyone goes home and never edits again. So you are not building a work group that can build and polish whole articles as a team. But ArbCom tends to destroy those groups when it finds them anyhow.

[Sitush]

The drawback of this kind of editing is of course that everyone goes home and never edits again. So you are not building a work group that can build and polish whole articles as a team. But ArbCom tends to destroy those groups when it finds them anyhow.

"Drawback" implies that the going home/never editing again is a result of GLAM. Is that really what you meant? FWIW, an oft-used phrase in the UK voluntary sector is "recruit, retain, reward". Two out of three wouldn't be bad; WMF barely manages one nowadays.

[Kelly Martin]

The drawback of this kind of editing is of course that everyone goes home and never edits again.

If these events do not lead to any recruitment of a least a few medium-term volunteers, then they are an abject failure, at least as far as a recruitment action.

I firmly believe the purpose of these events is to give the Wikimedians who made the arrangements a chance to network with whatever institution they're held at, so that when a position opens there they'll have a chance to escape the hell of working for Wikimedia and get a real job. Like so many things in the Wikimedia sphere, the real object of whatever is going on is to facilitate the movement of Wikimedia staffers to jobs that aren't at Wikimedia.

[sparkzilla]

if you were designing the WMF from scratch, how many people would you employ and how would you organise it?

Um. Just to let you know that we have been working on a better Wikipedia for news-based pages and biographies for the last year. For deeply academic topics an alternative approach may work, but any news-based article does not need experts to add sources.

1. Change the system away from being based on articles. This has the following advantages:
a. Better presentation of data. Data can be sorted and filtered and cross-referenced
b. Ownership of pages is impossible
c. Bias and harassment are lessened
d. System needs less rules, because focus is on the edit itself, not the person making it.

2. Assign random editors to assess edits. Make assessment of edits and fully anonymous for the person submitting and the assessor.

3. Implement revenue-sharing with the contributors

4. Have a single person, who has the ultimate vote on content decisions. Have a proper chain of command.

5. Have a simplified rule set. Remove or clarify ambiguous rules. Provide proper on-boarding for new users.

6. Employ as few people in the core organization as possible. Outsource everything.

7. Use the best technologies available. Mediawiki is 10 years old and it shows.

[When pigs fly]

Every editor gets fitted with an electric anal probe and everyone !votes to tickle their innards.

[Neotarf]

The drawback of this kind of editing is of course that everyone goes home and never edits again. So you are not building a work group that can build and polish whole articles as a team. But ArbCom tends to destroy those groups when it finds them anyhow.

"... FWIW, an oft-used phrase in the UK voluntary sector is "recruit, retain, reward".

So has Wikimedia Manchester discovered any magical formulas for that?

The drawback of this kind of editing is of course that everyone goes home and never edits again.

If these events do not lead to any recruitment of a least a few medium-term volunteers, then they are an abject failure, at least as far as a recruitment action.

I firmly believe the purpose of these events is to give the Wikimedians who made the arrangements a chance to network with whatever institution they're held at, so that when a position opens there they'll have a chance to escape the hell of working for Wikimedia and get a real job. Like so many things in the Wikimedia sphere, the real object of whatever is going on is to facilitate the movement of Wikimedia staffers to jobs that aren't at Wikimedia.

In my experience, it's the other way around. In the U.S. at least, for the most part Wikpedians are all volunteers, who already have established careers in the public or non-governmental sectors. Wikipedians facilitate the entry of WP into their own institutions, or institutions their own institution has a relationship with. And sometimes they balk, after they get a glimpse of the seamy underbelly of WP. Do they want to expose their institution to legal vulnerabilities over the obvious HR issues?