3 ms·

This is correct. It's an encyclopedia with few personalization features, so well over 90% of HTTP requests are served directly from cache. It may not be easy t

by neilk 15y ago

This is correct. It's an encyclopedia with few personalization features, so well over 90% of HTTP requests are served directly from cache. It may not be easy to explain this in a banner ad, but the money is really going to different efforts that are trying to expand / entrench / defend Wikipedia in its adolescent phase. Like: making operations run more smoothly, on more continents, and not require the constant intervention of a few wizards; enabling a virtualization framework which so more volunteers can work with ops or do wilder experiments (the "labs" project); inroads into academia; improving automated testing; expanding our ability to host multimedia; the GUI editor and usability projects; or the Wikipedia Zero project, where we try to get the site delivered to mobile devices in the developing world for free.

_delirium 15y ago
I do like some of those, especially the technical improvements to things like the editor interface, but as a long-time Wikipedian I confess I'm a little bit skeptical of some of the larger ambitions. I think Wikipedia's competitive advantage basically is writing an online, unencumbered encyclopedia, and providing useful data export so others can reuse and adapt it. When it gets into things like the Foundation itself improving education in the third world, there are a lot of complexities where I don't think it really has a competitive advantage over many of the other charities who have been working on that problem for a long time. So if I had my say (which I don't), I'd focus more on making the data more easily exportable in various forms so third parties can take on those missions, rather than Wikimedia itself becoming some kind of global NGO working to bring knowledge to the developing world (a mission that seems uncomfortably missionary / "white man's burden" ish to me in the first place). The situation has improved recently, but for years there was a bunch of hiring and spending on ambitious projects, but what I see as a core function of Wikipedia, the database dumps, which give the free-as-in-freedom practical effect (free licensing isn't that useful if you can't get the data) languished in a pretty poor state (with no full-history dump for a long time), which seems like misplaced priorities. Even today the dumps aren't all that user-friendly, even for technically inclined researchers, let alone less technically inclined adapters/remixers/reusers, and as far as I can tell, improving data export isn't one of the top Foundation priorities.
- neilk 15y ago
  I'm with you all the way on that one. But as you know, MediaWiki's brand of wikitext is famously impossible to work with. So, while it may not seem that way, really, the most important thing we're doing is redoing the parser and all the parsing infrastructure so it can export in forms that people can actually do something with. So if someone else wants to make the ultimate Wikipedia mobile app we won't have to be involved. Our official reason for redoing the parser is to make the GUI editor more possible, but IMO exporting is going to be the bigger application. In the meantime though, I still think the WMF has a role to play, particularly in making content available on mobile. Only this year, they finally got some of that straightened out (previously there was an inadequate Ruby on Rails proxy that was serving m.wikipedia.org). As a technologist, it's easy for me to point fingers, and say that some of the more ambitious stuff isn't warranted, or is too "soft", and we should focus on other stuff. But I have to say, my mind was completely changed by seeing how the academic outreach program worked out brilliantly. It will one day seem weird to have done a research paper that was reviewed and read by only one person. (Similarly, the people who work on that stuff have no idea what the "tech side" of the foundation are doing, in detail. Like if they ask when are they going to be able to upload more than 100MB long videos, it's hard to explain to them that we're undergoing a painful transition from NFS file storage to a media abstraction layer with a Python-powered distributed-hash-table + object storage server, and we're also making sure our new chunked uploading protocol works.) Anyway, it's a non-profit. These institutions are far from ideal and sometimes end up doing strange things for reasons which don't make sense outside the institutional context. The same things happen all the time at for-profits, but you don't get to debate it as much. :)

by neilk 15y ago

_delirium 15y ago

I do like some of those, especially the technical improvements to things like the editor interface, but as a long-time Wikipedian I confess I'm a little bit skeptical of some of the larger ambitions. I think Wikipedia's competitive advantage basically is writing an online, unencumbered encyclopedia, and providing useful data export so others can reuse and adapt it. When it gets into things like the Foundation itself improving education in the third world, there are a lot of complexities where I don't think it really has a competitive advantage over many of the other charities who have been working on that problem for a long time. So if I had my say (which I don't), I'd focus more on making the data more easily exportable in various forms so third parties can take on those missions, rather than Wikimedia itself becoming some kind of global NGO working to bring knowledge to the developing world (a mission that seems uncomfortably missionary / "white man's burden" ish to me in the first place). The situation has improved recently, but for years there was a bunch of hiring and spending on ambitious projects, but what I see as a core function of Wikipedia, the database dumps, which give the free-as-in-freedom practical effect (free licensing isn't that useful if you can't get the data) languished in a pretty poor state (with no full-history dump for a long time), which seems like misplaced priorities. Even today the dumps aren't all that user-friendly, even for technically inclined researchers, let alone less technically inclined adapters/remixers/reusers, and as far as I can tell, improving data export isn't one of the top Foundation priorities.

neilk 15y ago
I'm with you all the way on that one. But as you know, MediaWiki's brand of wikitext is famously impossible to work with. So, while it may not seem that way, really, the most important thing we're doing is redoing the parser and all the parsing infrastructure so it can export in forms that people can actually do something with. So if someone else wants to make the ultimate Wikipedia mobile app we won't have to be involved. Our official reason for redoing the parser is to make the GUI editor more possible, but IMO exporting is going to be the bigger application. In the meantime though, I still think the WMF has a role to play, particularly in making content available on mobile. Only this year, they finally got some of that straightened out (previously there was an inadequate Ruby on Rails proxy that was serving m.wikipedia.org). As a technologist, it's easy for me to point fingers, and say that some of the more ambitious stuff isn't warranted, or is too "soft", and we should focus on other stuff. But I have to say, my mind was completely changed by seeing how the academic outreach program worked out brilliantly. It will one day seem weird to have done a research paper that was reviewed and read by only one person. (Similarly, the people who work on that stuff have no idea what the "tech side" of the foundation are doing, in detail. Like if they ask when are they going to be able to upload more than 100MB long videos, it's hard to explain to them that we're undergoing a painful transition from NFS file storage to a media abstraction layer with a Python-powered distributed-hash-table + object storage server, and we're also making sure our new chunked uploading protocol works.) Anyway, it's a non-profit. These institutions are far from ideal and sometimes end up doing strange things for reasons which don't make sense outside the institutional context. The same things happen all the time at for-profits, but you don't get to debate it as much. :)