Thanks for your replies, and for John's very kind offer. As this involves various inactive and semi inactive bot operators I will give John an off list response re that.
On MzMcBride's points, I don't know the relative costs of employing programmers in San Francisco v other parts of the world. I am more familiar with the huge contrast in relative labour costs of London v Tbilisi. My point was that Python programming could be done pretty much anywhere so if we launched such a team we might as well do it where movement money would go furthest.
As for the idea that we want scalable, sustainable, and secure tools, I agree and think my suggestion would contribute to that. We still want volunteers to write bots that do useful things. Some of those will be deemed so useful and essential that they need to be incorporated into mediawiki, some will be transient things that might run for a few years but only be needed by their bot operator. Having a bot adoption resource would be useful for things that fit in between, ones where the wiki can live without them for a weekend whilst their bot operator needs their server for something else. But which if they don't run for months leave a loophole in our quality improvement programmes. So more scale able and sustainable than bots are today, but not as much as things that need to be added into the mediawiki code.
Regards
Jonathan Cardy
On 22 Feb 2015, at 21:04, wikimedia-l-request@lists.wikimedia.org wrote:
Today's Topics:
- Funding bot maintenance (WereSpielChequers)
- Re: Funding bot maintenance (John)
- Re: Funding bot maintenance (MZMcBride)
Message: 2 Date: Sun, 22 Feb 2015 11:53:53 -0500 From: John phoenixoverride@gmail.com To: Wikimedia Mailing List wikimedia-l@lists.wikimedia.org Subject: Re: [Wikimedia-l] Funding bot maintenance Message-ID: CAP-JHpm7g+FPosjye2=XH+iQYh8XCBPFb8Ee3tDSpKpdWKEhKw@mail.gmail.com Content-Type: text/plain; charset=UTF-8
I am a dev and am willing to replace a tool when it dies. I have a fairly large infrastructure of code that makes it fairly easy
On Sunday, February 22, 2015, WereSpielChequers werespielchequers@gmail.com wrote:
One of the areas that I would like to see the foundation putting in money is for the running and maintenance of wanted orphan bots. Wanted in the sense that editors are using them or would if they were still running, and orphan in the sense that the original developer isn't around or available to run them/migrate them to the latest platform.
If we work on the premise that community funds should go for things that volunteers want to have happen but aren't volunteering to do, then this is a classic and uncontentious niche. Programmers like to write new code and solve new problems, but the person with the idea or who writes new code doesn't always have the time and motivation to keep maintaining and running that code, let alone creating slightly bespoke version for scores of our thousand wikis.
Now it may be that we are in an unusual situation that the migration from toolserver to labs has cost us a number of bots that would otherwise have continued for years. But there will always be demand to localise existing bots for wikis where they don't currently run, and in the long run all of our volunteer bot writers are likely to move on.
Employing a python programmer or two somewhere cheap like India or South America would not be a huge investment for the foundation, but it would be a valuable service to the community, and unlike mediawiki development this could be completely volunteer driven with wikimedians deciding which bots are worth maintaining and their relative priority.
Disclosure: whilst I'm not pitching for the money for this, I would be front of the queue to ask such a maintainer to take on bots that I used to use the results of and in at least one case which I designed.
Regards
Jonathan/WereSpielChequers
Message: 6 Date: Sun, 22 Feb 2015 16:04:31 -0500 From: MZMcBride z@mzmcbride.com To: Wikimedia Mailing List wikimedia-l@lists.wikimedia.org Subject: Re: [Wikimedia-l] Funding bot maintenance Message-ID: D10FA52D.4D2B0%z@mzmcbride.com Content-Type: text/plain; charset="UTF-8"
WereSpielChequers wrote:
One of the areas that I would like to see the foundation putting in money is for the running and maintenance of wanted orphan bots.
I think specific examples might help here. If we're talking about category renaming bots or talk page archiving bots, I wouldn't mind if they died. The key is having suitable replacements in place first, of course.
Employing a python programmer or two somewhere cheap like India or South America would not be a huge investment for the foundation, but it would be a valuable service to the community, and unlike mediawiki development this could be completely volunteer driven with wikimedians deciding which bots are worth maintaining and their relative priority.
Do you have a ballpark estimate of how much money we're talking about per year per programmer? I'm mostly just curious how it would compare to hiring someone in San Francisco, for example.
Was the Wikimedia Foundation intended to be a technology company? Is the current Wikimedia Foundation suited to be a technology company or would it be better off contracting out development? These are probably higher level questions, but they're inter-related with what we're discussing here.
But more to your point about hiring cheaper labor, we don't know if a popular tool means that the approach taken was the best or should be sustained. We ideally want scalable, sustainable, and secure tools. I'm pretty wary of the idea that we could easily outsource this work.
MZMcBride
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines Wikimedia-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikimedia-l
End of Wikimedia-l Digest, Vol 131, Issue 48