Talk:Google hosting

Please think before you post! Many of the comments below are composed of rampant alarmist speculation bearing no relation to the facts. At a minimum, please skim the posts and responses already here first.

The Google_hosting page states:

The terms of the offer are currently being discussed by the board. The developer committee have been informed of some of the details via email. A private IRC meeting with Google is planned for March 2005.

Why are there no links to email archives or wiki pages with the discussions? What happened to radical transparency?

• Discussions by the board
• Details told to the developer committee
• The IRC meeting is already stated to be secret (private)

It is difficult to understand why the wikipedia board is unwilling to use wiki methods for such important type of decisions? i've spent so much time explaining to people why wikis are so useful - and i'm sure nearly all of us have the same belief in the power of using wikis - that i simply have difficulty understanding how people elected to the wikimedia foundation can forget what a wiki is all about. Boud 18:33, 20 Feb 2005 (UTC)

Google is a publicly traded company and is as such is somewhat legally limited in what can be said before deals are complete. There's also the question of respecting the privacy of donors who want their donations to be private or private until they are ready to make them known. Helping them to do that is the way to best facilitate those donations. Jamesday 07:20, 6 Mar 2005 (UTC)

Why have the two largest Wikipedias (English and German) been removed from the list of available downloads?

They are available for download.

Why is the last dump now 25 days old (February 9)?

Making the public dumps is resource intensive, we're still shuffling things around since the crash and need to set it up to run without breaking the site for a few hours at the peak parts of the backup. --brion 20:40, 6 Mar 2005 (UTC)

I don't know the details of the google offer as they're not public. For what I've understood:

1. Google Inc offers to give a couple servers

just say Thanks, take the servers and remove them from the to buy list.

1. Google Inc offers to give bandwith to wikipedia by hosting servers

just say thanks but no, google is a profit based company and profit is not part of the spirit of wikipedia.

Both work. You should also expect to see at least a second colocation site funded by donations from the general public in 2005. That is: it's not going to be an excessive reliance on Google, just people of good will helping eachother to do good things. Jamesday 07:20, 6 Mar 2005 (UTC)

Google's profits come from others who advertise on Google and other sites through them. Why shouldn't advertising revenues be used to support us? Brian j d | Why restrict HTML? | 05:56, 2005 Apr 3 (UTC)

Some are worried by Google Inc being on its way to global domination, (see google-watch) and other facts such as the fact that a majority (around 75%) of Internet search is made through google, gmail a powerful tool that can be misused to get information from a lifetime of emails, a google browser [1]: gbrowser [2] that can be misused to know what is your online behaviour, an OS integrated search engine that integrates with google internet search engine (a powerful tool that can be misused to get informations from inside your computer), numerous complaints about the adwords [3]. All of this tends to make some people nervous and suspicious about google. It is even more suspicious that google offers to host the wikipedia project which has for objective to compile all the human knowledge (which can be misused ... you got it by now). I'm not saying that google is not to be trusted, just trying to make a point that google is centralizing many powerful tools which could be misused in a terrible way for personal freedom and liberties.

As Google Inc is a profit based company, it is relevant to think that google plans on getting some profit from this offer to wikipedia, could be an image boost inside specific communities, or a long term plan for something else.

I agree that Google is centralizing many powerful tools but that just seems to make things more efficient. We won't have to wait for new servers to get our searches and we won't have to put up with that message: "their indexes may be out of date". Brian j d | Why restrict HTML? | 05:57, 2005 Apr 3 (UTC)

The simple solution is to require that google make the entire archive available to anyone for a nominal fee (covering their costs and overhead) for as long as Google uses the encyclopedia. If google agrees then their motives are most definitely altruistic. If not, they are almost certainly looking to co-opt the wikipedia archive and interface into google (similar to what they have done with usenet news, being the sole owner of a historical record).

Has anyone been able to work out what this means? What archive are Google meant to make available? I've been staring at this comment for a few days now and haven't been able to make any sense of it. Kate.

I'm wary of Google's access policy at the moment. I've been blocked from them because they - incorrectly - believe I've got an infected PC. I have no means of redress, and no contact details are given by Google.

This has nothing to do with your ISP having infected users. --Mboverload 03:23, 11 Feb 2005 (UTC)

Best make sure Google can't pull stunts like that if they don't like certain Wiki contributors etc. --202.49.141.2 01:43, 11 Feb 2005 (UTC)

I, for one, welcome our new Google overlords. --207.253.111.55 01:29, 11 Feb 2005 (UTC)

In Soviet Russia, Wikimedia hosts Google! Project2501a 09:07, 2005 Feb 11 (UTC)

If this works out, Google could be providing great help for Meta Wiki.

Will it be something like what Google Directory do with the Open Directory Project right now? .. a kind of "snapshot" page --81.155.140.162 01:39, 11 Feb 2005 (UTC)

If Google wants to host the Wikipedia content then Google should license the Wikipedia content since it was freely contributed and they will be making a profit on it's content. Income from that license should offset any Google overhead expense for hardware and management. Control over access and content must solely remain with Wikipedia via. contracted agreement protecting those rights with dispute resolution at Google's expense. - Daveanderson

This can mean one of two things:

1. Google sees a certain commercial promise (which is not too unclear if you think about it) in the wikimedia projects, and hopes to set its roots early. We've already seen that Google is making some very long-term investments (e.g. its recent dark-cable purchase, and employment of the plan9 designer). Also, Google undoubtly sees itself as the main - if not sole - dealer of information, in the forthcoming age of information trafficing. If this is the case (and only those who have seen the actual offer can say that), I'd say wikimedia should resist the offer.
2. Like most hi-tech corporation bosses, Google's founders are technocrats. They'd like to see society progress, and perhaps that's even the motivating force behind their Google enterprise. Like Ford and Carnegie, who have spent billions on universities, theatres and even political movements, the Google people want to help without looking for any profit in it. Wikipedia preserves technology, science, culture, arts - ideas. Preserving Wikipedia preserves human progress. And if this kind of thinking bred Google's offer, I'm all for it.

Just my two cents. 192.115.24.176 01:57, 11 Feb 2005 (UTC)

I'm surprised that the Internet Archive hasn't offered their resources. Wikipedia is a public resource that might suffer from any commercial involvement.

(Please sign posts.) The Internet Archive's resources aren't any better than Wikimedia's. They took the Wayback Archive search function offline a few months back, and the actual archiving is extremely spotty today, even at sites that were getting daily blinks back in '01 or so. So, don't think of them as a potential savior. They have nothing like the potential bandwidth and server resources that Google has "just lying around". --68.249.74.118 06:18, 12 Feb 2005 (UTC)

It would be impossible for Google to do anything that would adversly affect wikipedia. Since wikipedia is community run, they cannot complain/do anything about about an article citing bad things google has done. All they are doing is offering hosting and a more integrated search.

Agreed. Remember, all the content is FDLed. Google cannot co-opt it, and there is nothing for them to "own." It seems to me that the best they can do (as others have pointed out) is indirectly benefit from a working Wikipedia. Even worries about ads don't make any sense, as no formal organizational agreement with Google is needed to utilize AdSense, so I don't believe this hosting arrangement has anything to do with in-page ads. --Akrowne 22:53, 13 Feb 2005 (UTC)

This is not time for the tinfoil hats. Put them away, save them for the when the MPAA sues Wikipedia for having articles about movies. --Mboverload 03:27, 11 Feb 2005 (UTC)

These guys are already copy catting microsoft! The only reason Google wants an encyclopedia is because the new msn search can search through encarta (pretty sweet stuff). I guess google with their billions doesn't want to pony up for their own commercial encyclopedia!

That's stupid. You can't buy an encyclopedia like Wikipedia with cash. --wikt:en:eean

Right, that's why they're offering bandwidth instead. ;-) --GreenReaper

Also, google already searches wikipedia and its ad-sponsored mirrors. Can't be it. 192.115.24.176 12:16, 11 Feb 2005 (UTC)

Folks.. not sure if this has been precisely mentioned before, but just wanted to chime in. As long as the entire MySQL database (I think that's the format everything's stored in..?) of all Wikipedia articles, images, etc. remains publicly and freely available, either from Google itself, or from a wikimedia mirror, Google can do nothing to restrict access to Wikipedia's knowledge, either now or in the future. Everything's licensed under the GFDL, there's no way that Google could steal the copyrights on the article (since the copyrights are owned by the individual contributors, but released under a GFDL license).. if they tried to put ads on anything, or charge for access, or somehow restrict access, we could just go back to hosting it on Wikipedia's servers. Problem solved. I see little risk in letting Google act as a mirror.

As a matter of fact.. now that I think about it, nothing is stopping Google from downloading the entire Wikipedia archive, installing the Wikimedia software (the wiki backend), and just having their own local mirror of everything, but without being an "official" mirror or whatever it is that they're proposing. Schmiddy 20:13, 15 Feb 2005 (UTC)

We don't actually make the whole database available. We skip the account details (email addresses, passwords and such) as well as some recreatable tables. I'm one of those who produces the downloadable database dumps and I want to ensure that thre's enough in them so that the site getting hit by a meteor tomorrow won't lose the information. Google is at liberty to publish their own copy of the whole Wikipedia with ads if they want, just as anyone else can and some do. But no reason for them to do that when we can be specialists in that area. Jamesday 07:20, 6 Mar 2005 (UTC)

So the money could be spent developing open technology for the site. I doubt the only thing that's needed is just bandwidth and CPUs. 80.111.239.225

The vast, vast majority of what we need is hardware. I'd rather have more hardware than less money. We do have our own money; if you took out the hardware costs, it'd pay for everything we need several times over. Kate.

Wikipedia is the top result for a growing number of search terms in Google, and on the first page for many more. The site can't handle all that traffic, so lots of people end up having to click on the "cached" link after waiting a few seconds and seeing an error message. For them, Google is *already* hosting Wikipedia. This sounds like nothing more than a way of making it official, and saving a few seconds for people who click on Google's Wikipedia links.

Excellent point. I would add the following: since Wikipedia does come high up for so many of Google's results, the smooth running of the Wikipedia site is in Google's interest - Google's value to its users comes from the fact they can use it to get the answers they want quickly. So if Google spends a few 10,000ドルs on ensuring Wikipedia rules smoothly by hosting it themselves, it will easily pay for itself in terms of user satisfaction for Google itself. One caveat: the agreement has to be permanent. Wikipedia can't get itself into a situation where it's become dependent on Google, and then Google goes and changes the hosting terms, and Wikipedia has no choice but to acquiesce because it has no other options.

I'm not sure if we are not shooting ourselves in the foot here -- the agreement shouln't permanent; if it is, then we are dependent on Google. Nothing lasts forever, and signing a permanent deal is the fastest way to become dependent. ✏ Sverdrup 17:08, 11 Feb 2005 (UTC)

I don't understand what this means. How would Google offering hosting "permanently" mean that we are automatically dependent on them? Kate.

If Google suddenly 'went bad' after Wikipedia had committed to a 'permanent' hosting of the central servers at Google, this could make life really hard on support and admin, etc. On the other hand, without some sort of long-term committment (e.g. a 1 year transition period in case of a falling out), Google could, as implied above, treaten to pull all bandwidth and hosting 'next month' as a gun to the society's head.

If Wikipedia doesn't have to provide it's own hosting for a long time, it could take quite a while to regenerate the infrastructure to do that again, in the case of a falling out. The longer the free-fall, the harder the impact (modulo terminal velocity). Parachutes are good safety features. -- [darkonc]

I also agree with this point. All that will probably happen is that Google hosts a mirror that Google search results seemlessly links to (ie no funny URLs). Any other search engine should be given the same option. If the big engines all did this, the central serving requirements would be greatly reduced.

Another likely option is for Google to treat the Wikipedia like it does the Usenet - a freely available source of free data it can add features to, so that the majority accesses the data through it (and past their ads). They attempted something similar with the Open Directory, but were less successful. Some say Firefox is next.

IMHO they were quite successful with the ODP. "Google Directory" is better known today than the ODP itself.

Such an approach could result in the "Googlepedia"; and as with IRC and mIRC, the name "Wikipedia" may never be known by the majority that use it. Matt Cook

Which is quite possible, but, of course, completely unrelated to the fact that Google have offered to host some servers for Wikimedia. Kate.

sounds like a great offer. could google also sponsor a decent search? eg. google enterprise search

40 edits, and all that's been achieved (apart from vandalism by /. twits), is [4] changing from plural-use of a group of people (Google) to singular - "Google bave" -> "Google has", a BE -> AE change.

Does meta have a policy against such changes?

James F. (talk) 11:16, 11 Feb 2005 (UTC)

Yes, the policy is at Official Meta Policy on Spelling. ;) Angela 11:25, 11 Feb 2005 (UTC)

Don't know about BE to AE in this case - I'd write has because I'd be writing about a single company. I'd use have if I was writing about the employees of Google. Jamesday 07:20, 6 Mar 2005 (UTC)

The points 1 and 2 above are good. How about we assume point 2 (altruism or at least something non-harmful to WP), and protect the rights of WP and its content from 1 (commercialisation). I couldn't care less what people's motives for helping WP are (commercial, altruistic, strategic)... so long as it does actually help (ie no rights lost, no sneaky takeovers or anything). Hosting and such from google sounds pretty synergic to me.

Let us assume 192.115.24.176's second option: that Google is treating this as Bill Gates does a library, as a humanitarian act. Unlike a donation such as a batch of hardware, Google is offering an ongoing service. If Google ever goes bankrupt or gets bought by a less generous company (both unpleasant but fully possible events), the resources they donate could be snatched up and repurposed, leaving Wikipedia dry and gasping. While I don't see this as a complete argument against the agreement, it's a risk all parties have surely taken into account. NickDouglas 13:32, 11 Feb 2005 (UTC)

This is a known risk, which is why the proposed hosting, if the deal goes through, would be only one of several sources of hosting. Today we rent space in a data center in Florida and have hosting donated for several additional servers in France; these would not disappear. --brion 06:17, 15 Feb 2005 (UTC)

Don't assume centralization is bad. If Google were the only host, their search engine could give up to date results when searching Wikipedia. I presume this will not be the case the way we are going now. Brian j d | Why restrict HTML? | 03:00, 2005 Apr 7 (UTC)

This seems like pretty basic math here. The more useful info out there on the net, the more useful google is. the more useful google is, the more money they make. wikipedia has proven to be a very successful way of creating a huge amount of content. it also does things (and could do more things) that google (and other search engines) might be able to use to make their search results even more relevant. specifically things like categorisation and timelines and multiple languages could be a great help to search engine authors.

it might be a good idea for the wikipedia team to see if google would be willing to make document any mediawiki/google interfaces that improve searching - in the interests of helping other search engines (good for competition) and other content systems (good for sites, google and its users). not the algorithms google uses obviously, but the interfaces and how they might be used.

likewise it's a win for wikipedia. there's bandwidth gained for sure, but it wouldn't be a bad thing for mediawiki developers to work with search engine authors so that each can get a better understanding of their issues.

i think it can be a win for google, a win for wikipedia, a win for other search engines and a win for users. there is lots of room for growth - billions of people have never used the internet. everyone can come away with something. skepticism and realism are good things and i hope people think through the consequences. but likewise hopes and goals are good things too and i hope they are given consideration as well.

--KevinLyda 14:34, 11 Feb 2005 (UTC)

It's better to rationally list the pro and contra of such a move. In my experience

1. You never "give away" content in exchange of Communication Capacity (all that is needed to copymove a bit from point A to B) 2. Because the CCapacity is relatively cheap when bought in enormous quantities , like Google does 3. The Wikipedia Brand, Wikipedia Database, Wikipedia Userbase, Wikipedia Concept are entities who raise the value of otherwise almost valueless CCapacity

Google would benefit noticeably by adding Wikipedia results to their search results a) by increasing satisfaction of users (offer them more kind of results, let them pick their choice) with an almost 0 cost increase b) by captivating user attention for more time (incresing likelyhood user sees a commercial) c) by selling sponsored links further sponsored by Wikipedia co-branding d) by slapping another interface on the content and sell adds while destroying Wipedia concept and brand

In exchange, Wikipedia would receive theoretically infinite bandwidth and server capacity, but not necessarily quality of service neither service : that is to say, Google isn't (afaik) assuring Wikipedia free-users (not paying) they're going to receive any access or timely access to their database of choice (Wikipedia) nor that edits of the database and backups are made promptly avaiable for consultation and/or modification.

Imho, it is best to remain highly mobile and not to give up either content and/or brand..the moment you become dependant is the moment they screw you (not the Google tech guys, those are usually nice grunts...but the Google corpomotherfu guys, those are more like leeches) --84.220.22.101 14:54, 11 Feb 2005 (UTC)

There are no plans to give up either content or the brand. Google (and Yahoo and MSN and heck probably Al Qaida) can copy our content and slap ads on it all they want, this is the case already and is unrelated to the proposed possibility of providing some additional server hosting. --brion 06:11, 15 Feb 2005 (UTC)

A consideration is that you are selling the privacy of your users. Google actively data mines their usage logs, and their pack of data miners will no doubt be chomping at the bit to analyze this 'wiki' phenomenon once they have your logs.

The immediate advantages to Google is simply reducing the time delay between indexing Wikipedia and reselling that information to their customers. In practice, they could just mirror Wikipedia themselves without you, but that is too slow. The parallels to Usenet I think are the clearest. It's faster for them to be the main source of information than to have to mirror it. The concern should be about the long term benefits to Google that disadvantage Wikipedia.

Since Google is infamously secretive and therefore untrustworthy from a non-profit organization's point of view (not saying anything about the people, just the structure of the situation), so it's best to take a cautious stance towards them. What specifically do they want (disbelieve altruism--Google is not altruistic), and what is the minimal amount of 'help' they can offer in exchange for you helping them? If they want real-time access to the data, you can sell preferred access to your server's data without selling control over the server. If you did this wisely, you can also sell to other people who want dedicated pipes to your server and actually generate lasting sustainable income. Suppose Google reneges? Their competitors will keep them in line, or at least protect your income if one of your customers drops you as you will more sources of income.

In principle, a negotiated win-win relationship, where it is obvious what both party are benefiting, is more stable and less risky and more predictable than an altruistic relationship since there is nothing guaranteeing that altruism. -- Sunir Shah

Getting updates from the wikis would not be particularly related to the possible physical location of some servers which we operate in one of Google's data centers. That would be an entirely separate issue both in principle and technically. --brion 06:07, 15 Feb 2005 (UTC)

The access logs we have now are private and I assume that that will not change in this case. I also assume that we'd be the people doing the day to day server operations, with Google effectively being a colo. Google probably already has quite a bit of access data available from knowledge of which links people click on on their own site to get to the wiki content. Jamesday 07:20, 6 Mar 2005 (UTC)

Is anyone worried about what the influx of newbies to Wikipedia from Google might mean? Imagine if Google decided to add a Wikipedia search to the to of the homepage alongside Froogle and the like. Suddenly you'd have many, many more people poking around that don't really know what they're doing. I think one of the biggest reasons wiki's have been so successful is that the user base is primarily those who actively seek resources like this out, rather than those who stumbled upon the site by accident. Parnering with Google could substanitally change the demographics around here. Whether that's good or bad shouold be discussed.

We already get a huge amount of traffic from Google searches. --brion 21:20, 14 Feb 2005 (UTC)

We're already huge. Top 75 English language sites on the net, top 125 worldwide according to Alexa. According to Alexa now more popular than the Altavista search engine. More than a million pages. Jamesday 07:20, 6 Mar 2005 (UTC)

Since the entire content of the wikipedia is available for download already, there's no reason they couldn't host their own wikipedia content, pull it from us, and do whatever indexing they want to do with it. If they have some altruistic goal, they should instead just donate money. They've certainly got the spare cash. -- Avriette Sat Feb 12 00:51:34 UTC 2005

Google buys bandwidth and servers in ridiculous bulk quantities. Directly offering hosting is far cheaper to them than donating money for hosting somewhere else. Altruism only goes so far.

The database being downloadable as a whole means Google can always have a current snapshot of wikipedia to set up their own fork. Google could easily integrate wikipedia into their interface (ie, just a front end for the official wikipedia) and, at the point where more people are editing content using their interface, fork the database and completely marginalize wikipedia. They could probably pull off a fork directly except for the negative PR it would generate and the lack of an existing community of contributors.

The license terms do not require making the entire database available, so there's no reason requiring Google to let people download the future changes. If they do not make it available then Google has an automatic exclusive lock on the future state of the 'pedia archive. That's not good, since people will stop contributing to the "official" wikipedia as it will be out of date, the donations will dry up, and the organization will fail. That's why I suggested a license term for Google requiring them to make the entire archive available for a nominal fee for as long as they make wiki content available.

Such a license does not sound like a free/open content license, and as changing Wikipedia's license at this point would take a very long time (75+ years), it's not a particularly relevant suggestion anyway. Kate.

Then google has no ability to lock up the archive (as a whole entity) behind their servers, because anybody (Microsoft or Yahoo) for instance has access to the data. But make sure the cost of them breaching that contract is high enough, for example if it is a 1ドル million fine for not making the archive available then they could easily take that hit to control the data.

Frankly I would rather solve the hardware and hosting issues by more strongly appealing for donations.

Non sequitur. Everything you said above is completely unrelated to whether Google can fork Wikipedia/"lock up the content"/do anything else. Providing hosting for us does not make it noticably easier for them to do that, and certainly does not mean they can "control the data" or anything else. Kate.

Every so often NPR and PBS really annoy their patrons in order to raise money, but everybody understands that they have to. If millions of people start visiting wikipedia from google and it slows down then donations should go up considerably -- especially if you really ask for them. Please consider this deal very carefully.

Relying on donations is fine, but given the growth of our projects, there is a limit to realistic expectations. We need partners, they can be organisations like Google, Kennisnet name them. Partnering should mean a win win situation. Money is not the only argument, culture is another. Given its behaviour Microsoft is incompatible. Google is not.

One factor that will save Google money by hosting us is that they do not need to keep a copy of our projects in their cache. Their cache IS the Wikipedia content. I would not be too suprised if they use more servers than we do to keep our content available on line. Another benefit to us will be when Google people will be interested in the optimalisation of our content and this way increase the number of developer working towards an even more grandiose webexperience than that we already provide. GerardM 11:18, 12 Feb 2005 (UTC)

Answers.com is already regularly updating their search engine with Wikipedia content//85.226.10.174 21:27, 12 Feb 2005 (UTC)

I would agree that this is not a decision that we should take lightly, particularly if it is an agreement that could not be replicated "in parallel" with multiple different search engines. We should not be swayed by the enormously good PR google have built up over the years to give then something we would not equally happily give to MSN or Yahoo. Wikipedia's strength lies in its independence. If Wikipedia must form partnerships in order to survive, I would suggest that it would be better turning to Universities for help. The philosophy department of my University has just been hit with a request to give 3000ドル to The Stanford Encyclopedia of Philosophy in order to keep it free (and ad free), which the librarian says she feels a "moral obligation" to give them. In this manner the encyclopedia expects to raise 3ドル million. (See their fundraising page). I have no doubt that if Wikipedia attempted a similar thing, it would meet with equal success. (I would even venture to suggest that UK Research Councils would be happy to give money to it.) In almost all democratic countries there is a constitutional convention that says that one parliament (or similar governing body) cannot bind the actions of a future one. To accept the Google offer would seem like doing exactly this; it is leading us down a road it will be very difficult to return from. I am glad that the Google offer has been made, it is good to have as a backup in future in case funding really does dry up, but I do not think we have yet fully explored all the options available to us. Accepting the Google offer would seem like a very un-wiki thing to do. The wiki solution is to get users as involved with fundraising as they are with contributing. --cfp 18:31, 12 Feb 2005 (UTC)

You seem to be making a lot of assumptions about what is on the table. If hosting from Google is accepted, it would be in the form of some additional servers at one of Google's data centers being put at our disposal. This would be in addition to and not replacing our existing and future hosting elsewhere, nor would it give Google editorial control of the projects or anything else that would compromise the Wikimedia foundation's goals. --brion 06:02, 15 Feb 2005 (UTC)

I truly hope you're right.

No-one is "right" or "wrong" about the terms of the agreement because it has not been agreed yet. However Brion is correct in the offer that Google has made; this is a plain and simple fact. Google have offered servers and bandwidth. Nothing else is happening. Providing hosting does not give someone control over the content. It is not "un-wiki" to accept donations of assets other than money; where do you get that idea from? Google are not the first people doing this and won't be the last. Kate.

If you are, may I suggest that to proove our search engine neutrality, we contact MS, Yahoo, FAST etc. and say that if they ever wish to make us a generous offer on the exact same grounds as google, then we will accept it. Of course they probably wouldn't want to, but I feel it'd be a good proof that we weren't "getting into bed with google" in any way. --cfp 15:57, 15 Feb 2005 (UTC)

If someone wants to code a "preferred alternative search engine" user option which works with a wide range of alternatives, that would be welcome. I'm usually the person who decides if and how much of search goes to the alternative search engines. I don't consider how much money it'll make them when I do it, even though I know it wll make the money. I decide based on our own server loads and a balance between speed and number of local searches - search is more than half of the database server load and we have an alternative, so it' a good option when load isuncomfortably high. Expect to see at least one more physical location financed by donations from the general public this year, in addition to anythign involving Google. We want that for reliability (so a problem in Florida is less problematic) and to help provide a better service now we have the resources to use equipment in ways which don't use every last bit of available performance as efficiently as possible. Jamesday 07:20, 6 Mar 2005 (UTC)

Public information currently is:

• A member of the Wikimedia Foundation Board of Directors has said that Mr Wales met with two Google representatives on February 10. [5]
• The Foundation's treasurer has given some details [6] about some possible outcomes, although these remain unconfirmed.
• Mainstream media, ranging from the New York Times to Slashdot, has been reporting rumors of such a deal as well.

I'm not sure why the attempt to suppress information. The multiple reverts by Wikimedia's president with no edit summaries are also particularly odd, and would be bannable were it anyone else. Isn't this still a wiki? --Delirium 07:37, 12 Feb 2005 (UTC)

The treasurer did no such thing. Certainly not in any official capacity that would warrant the use of a title. This is not a place for 'public information' since it is being referred to in the press specifically as an official statement by foundation. So only official statements go here. --Daniel Mayer 10:52, 12 Feb 2005 (UTC)

A meeting on February 10 was mentioned on the mailing list - but there's nothing which necessarily ties that to the hosting. It could have been a completely unrelated meeting afaik.

mav is not the treasurer, and hasn't seen the proposal.

Angela 12:12, 12 Feb 2005 (UTC)

An admin should correct the spelling of "responsability", as quickly as possible if the press is referring to this page as much as is said. Taco Deposit | Talk-o Deposit 17:40, Feb 12, 2005 (UTC)

I've removed the "responsability" part. Angela 19:47, 12 Feb 2005 (UTC)

Please remove the comma from "March, 2005." The proper form is "March 2005," without the comma. Neutrality 16:44, 14 Feb 2005 (UTC)

John Dvorak comments on this proposal here.-202.147.43.82 23:51, 14 Feb 2005 (UTC)

Unintelligently comments on this proposal, that is. He never once mentions the fact that Wikipedia is under a copyleft license. That's an important detail when you're trying to stir up fear that Google could somehow take over Wikipedia. I can't take Dvorak's worries seriously unless he acknowledges and adresses the licensing issue. Isomorphic 22:05, 16 Feb 2005 (UTC)

Ross Mayfield (CEO of Socialtext) responds to Dvorak's comment here -- Arne (akl) 09:48, 17 Feb 2005 (UTC)

This project really could use the speed and reliability of Google's servers. But the gatekeeper power the company would acquire over us is quite daunting. - Davodd | Talk 03:49, 15 Feb 2005 (UTC)

Since Google would only provide a portion of our hosting if a deal is made, we will always be able to do without them. Complete dependence would not be an acceptable condition. --brion 20:57, 15 Feb 2005 (UTC)

Often when I search in google, google gives me a mirror commercial site of wikipedia instead of giving wikipedia article itself. The mirror site is not up to date, the graphics are terrible, you cannot edit it etc . If the proposal can solve that, it will be great.
Besides, not all wikipedia pages are indexed by google, so actually the web surfer has no way to get some article. As a very active contributer in the Hebrew wiki, I find it very disappointing to see that article for which you have devoted much time and thinking isn't available to his potential readers. I suppose that the Google's proposal can solve those problems. SHASHAZ

I don't see how the physical presence of some of our servers in one data center or another would be related to the relative placement of sites in search result listings. How do you think this would be so? --brion 20:54, 15 Feb 2005 (UTC)

The problem is not the relative placement of sites. Google gives a mirror site when it doesn't know a wikipedia article exists at all.

Do you have an authorative reference for this assertion? Kate.

It seems that its crawler doesn't get everywhere sometimes. I suppose the indexing will be made in more efficient way if wikipedia is on google servers. SHASHAZ

I find this doubtful. While Google could perhaps point their crawler at specific locations, this is likely to involve hostnames, not IP address blocks. They could (or could not, depending on their setup) prioritize Wikipedia *completely* independently of where some of our servers may or may not be physically located. --brion 21:27, 15 Feb 2005 (UTC)

Yes, I do think that this will help Google to have a more complete index. Chances are that they ask for the next page only after the prvious one has been delivered, so making us faster would help them to have more complete coverage. That applies to all crawlers, of course, not just those of Google. Jamesday 07:20, 6 Mar 2005 (UTC)

Though getting server space and bandwidth on the margins of one of the Internet's largest companies may seem like a good idea on the surface, Google remains a profit-oriented business, and in the end that will not bode well for Wikipedia: They will find a way to make money off of the program somehow. They might actually be able to kill it off or fork it on a whim as well. The worst IS possible.

Instead, Wikimedia should pursue additional servers and bandwidth from institutions such as Universities, whose ideals are more inline with that of Wikimedia, and have significant bandwidth that won't vary with the business cycle. -- Alexwcovington, 21:04, 15 Feb 2005 (UTC)

Google, like everybody else, can fork any of our projects at their whim. This bears no relation to whether we have some (not all!) of our machines hosted in one of their data centers or not.

If a university is seriously interested in providing some additional hosting as well, let them make an offer. --brion 21:29, 15 Feb 2005 (UTC)

I've rarely (if ever) seen a public company acting out of altruism. Google is one cool company -- sure. They're cool with their employees, their technology is cool, they compete with MS and all -- but they are still a publicly held company. The shareholders do not necessarily share the same goals as the Wikipedia community.

If Google wants to donate to our project, I'm OK with that. But I don't want them to profit (i.e.: the shareholders get more money) from our work in a more direct way.

There are many ways Google (or any other person or corporation) could profit directly from our work without giving back a dime or an ounce of effort to the project. This is because we've chosen to release our work under a license which allows people to redistribute and modify our work. While there are restrictions in that license, it does not require redistributors to for instance avoid making a profit or remit a portion of profits to the original authors.

This of course is completely unrelated to the proposal that Google might provide the project with some additional servers and hosting (which would be only part of our hosting; we already have servers in two datacenters, one in Florida where we rent space, another in France where some space is donated to us). --brion 02:08, 18 Feb 2005 (UTC)

It must be noted that Google is offering to host Wikimedia content, not pay for it. By hosting it, Google automatically gets some technical control over it. Google's aim might not be to destroy Wikimedia, but, as a company, Google looks after itself before everything else. And this might not be a good thing. If Google simply pays Wikimedia for hosting, then Wikimedia might consider it. Anything else and I would prefer Wikimedia to reject it. But if they want to go ahead anyway, at least Wikimedia retaining complete technical control over the servers would be appropriate. Any more control to Google, including administrative control, would definitely be bad. Even if Google is just hosting the servers and is letting Wikimedia have control over the servers, you still can't tell what Google might try. It might tell its engineers who maintain the servers to do something which is simply dirty (for the benifit of Google, of course). Yes, this sounds a bit paranoid, but things like this have happened before and it'd be wrong to proceed without at least considering this.

If this does happen, then Google will only host a part of our traffic. We should never be too dependent on any one hosting partner (except for Wikimedia, of course - but we control that organization). The people who run the Wikimedia Foundation are well aware of the potential pitfalls of this type of deal with any potential hosting partner. Rest assured they will not depend too heavily on any one (at least for long). --Daniel Mayer 03:56, 1 Mar 2005 (UTC)

Maybe I see things too simple, but in my opinion google seeks to be the chore of all knowledge, the site everyone's going too if they want to find something (something like 50 percent of all searches on the web are made from google). That's why they put thousands of books online for free. Google sees wikipedia as a potential threat in the long therm, because more and more people go directly to wikipedia if they want to have information on a subject. So they want to incorporate wikipedia into their company. They'll want to keep it just as free and accesible as it is now, because that is the best way to stimulate growth, and thus strengthening more and more google. It's just a move with relatively low costs to improve their centre of all knowledge image. So they win, wiki wins. What's the problem?

Don't forget that wikipedia uses google as an internal search engine, and so, google knows everything about what people are searching for on wikipedia, and what keywords are used. From a project point of view, this fact is considered a major security problem , in terms of economic espionage and industrial spying. Keywords are the keys of web ranking and web ranking professionals are willing to pay for such information. To gain access to these information is the main objective in the search engine war we can witness between microsoft and google. For these reasons private companies usually ask for their own search engine for their website. Google had these informations about wikipedia, and also knows of the traffic and servers difficulties of wikipedia. As google is a profit based company, it makes sense to think that although not evident for us, they already know how to make profit with such an offer.

I had an idea for your deal with Google. I know you say there are not asking for anything in order to ofer the hosting, but just in case. ¿How about giving to Google, in counterpart for anything he wants to give you or you want to ask him for, the possibility for users and editors of wikipedia to have the choice of associating their account with a google address of their own and then having the possibility of letting google use the information about all or some of their going abouts in the wikipedia? Then Google could use that information to: -think about wikiusers, create nets of words related to people for the sake of using the net of words on its own, extract useful information about people in general, and different intesest people. -to invent services for those users (maybe in a wiki way form too) (NOTE: could it be dangerous that goolge learned to work as a wiki?), wether in a commercially direct and eficient manner for them or a more future-hopeful way. And essentially from gmail, not from wikipedia use (unless desired by user and "not evil"). A good idea I read was to try/ask to get another offer on the same grounds. Hope it is of some use or interest. Pablo Segundo (en.wikipedia lover and user from Spain) pablo2garcia at g

Google is patented. Or to say this a bit more carefully, it uses a patented algorithm. See patent.

So?

Google is now a billion dollar corporation. It will (presumably) sue the hell out of any group that tries to use any vaguely similar algorithm - if that group is in the USA, anyway. (But i'd like someone to prove me wrong.) Patents on software are not (yet) valid in EU law, but that doesn't mean that anyone is willing to take the risk in practice...

So?

AFAIK, wikipedia does not threaten anyone redistributing wikipedia content, provided that the conditions of the GFDL are adhered to, of course.

So?

IMHO, wikipedia ought to be extremely careful about being constrained by a group supporting software patents... And i agree with the arguments above about google being constrained by the market, whose interests are not those of human rights of ordinary people, except according to the rather confused (IMHO) ideas of objectivism, apparently beloved and believed by Our Leader, Hallowed be His Name. Boud 03:27, 1 Mar 2005 (UTC)

Your wiki is definitely not mine. I don't know what you're doing, but I'm building a free content encyclopedia. Stress "free content". The rest is bollocks. I couldn't care less if Google's shareholders ate babies, and their CEOs supplied them with said infants. "Google uses patents, so Google is evil, and we shouldn't deal with evil companies" is, to put it mildly, an utterly asinine way of reasoning. For the millionth time: Wikipedia is not going to haul its entire database over to Google as its sole bandwidth provider and delete all the mirrors. Jimbo Wales is not going to spend the fundraiser donations in Google stock options. We are not going to burn effigies of Richard Stallman under the portrait of a smiling Bill Gates. We are not selling our soul to the devil, here, much less our content. This "Live Free or Die" crusade nonsense is really irking me. 131.155.71.93 12:54, 1 Mar 2005 (UTC)

Let me clarify the discussion here: The wiki is both Boud's and 131.155.71.93, and the point here is to know that wikipedia provides freedom fo speech, and support open source whereas google provides censorship (see the example of google in china) and support the software patenting. As these are mutually exclusive ways of existence, this fact should be kept in mind along the discussion about the google offer, and indicates the need to be very careful in the process of reflexion before answering google. 82.66.65.136 16:20, 3 Mar 2005 (UTC)

I'm sorry if my previous comments may have fallen into the "rampant alarmist speculation" category, I hope they did not though. I am now more comfortable with the idea, particularly in light of brion's comments.

However I think the former worries I shared with others are illustrative of the fact that Wikimedia must very carefully manage how they present any deal with Google to the public. Certainly it will be important to make clear that it is truly a donation (no adverts, control etc.), but equally, it will be important to make it clear that there is no degree of exclusitivity implied by the deal: our relation with Google will be an "open" one. If anyone else wants to "get into bed" with us (donate bandwidth), then we'll just as happily accept it from them, even if they're a direct competitor of Google. I could continue the relationship metaphor, but I'm sure you can see where it would lead... I hope this is reasonable, and does not fall into the "alarmist speculation" category. --cfp 18:25, 1 Mar 2005 (UTC)

I can't find a flaw in Google's offer to host servers -- I wouldn't find a flaw in a similar offer from Microsoft (would you?). My only concern is how this deal is progressing.

• Discussions about the proposal are private, and cannot be made public because Google is publicly traded.
• The page is protected, because of vandalism a month ago. But edits by respected users with additional information are reverted.

John Dvorak's blog is paranoid nonsense not only because Google merely wants to donate a couple of servers that Wikipedia won't be dependent on, but also because Wikimedia has a culture of openness whose members would rebel against any imposition of control by Google even if it had the power. But IMO, it's precisely the erosion of Wikimedia's culture of openness that is the issue! It may be that there is really no more secrecy here than with any other hardware donation -- or it may be that we just can't interact with publicly traded companies if we want to remain as open as usual (I don't know which, if either). It remains true that most Wikimedias cannot edit this page -- and I (who have that power) am scared to use it, since that would mean reverting Jimbo.

Incidentally, having read a fair amount of ill informed conspiracy theories just now (mostly not on this talk page), let me offer one that at least has relevance to Wikipedia: Google has no intention of touching the Wikipedia proper, where the free (as in beer) editing is done. But they want in on the ground floor of the new Nupedia, aka Filterpedia, aka Wikipedia 1.0 (sans the offline aspects). Spread that one around and look savvier ^_^.

-- Toby Bartels 16:31, 21 Mar 2005 (UTC)

Now is april, is there something new information? 80.185.26.193 11:10, 2 Apr 2005 (UTC)

Soon, allegedly... Patience! :) --brion 00:35, 8 Apr 2005 (UTC)

Any news at the end of April? Why don't you just tell us? 158.193.85.212 15:36, 27 Apr 2005 (UTC)

No new news :-) but other news coming... Anthere

Still no news, but I've updated the page to explain why there's no news. Angela 19:13, 28 May 2005 (UTC) Reply

According to the protection log, this page was protected nearly two months ago, to prevent Slashdot troll vandalism. I'd like to request that it be unprotected now, so that users can add other relevant links and information. -- Creidieki 21:26, 7 Apr 2005 (UTC)

Indeed, it needs to be updated. Since the google thing has been officially announced and set now, I've unprotected it. TUF-KAT 21:49, 7 Apr 2005 (UTC)

Oops, my mistake. That was a deal with yahoo that was announced, not google. I've re-protected it, since presumably whatever reasons it was originally protected for still apply. I'm not saying I believe it should remain protected, since this issue is no longer extremely prominent, but I'll wait for consensus from others since it hasn't been announced. TUF-KAT 21:51, 7 Apr 2005 (UTC)

I've unprotected it. Angela 19:13, 28 May 2005 (UTC) Reply

Which april? :) It's gone [april, not google], so either the page should mention that the irc meeting failed/cancelled, or tell what happened, and what's next. This may well point back to "page needs updating" requests, but I tried to phrase it to help people what I mean needs to be updated. :) --grin 10:03, 5 May 2005 (UTC) Reply

There never was a meeting. The page is now updated, and explains why nothing has happened just yet. Angela 19:13, 28 May 2005 (UTC) Reply

"We are still talking with Google about creative ways in which they might support the Wikimedia Foundation. The current status is that we are digesting several hosting offers at the same time, and it would be premature at this juncture to ask Google for something specific when we don't yet have good technical knowledge of what we need. They are eager to help us, we are eager to accept their help, but we want to be good stewards of donor money, and this requires us to move carefully.

The Yahoo facility is not online yet, the facilities in Amsterdam and Belgium are not online yet. These things take time to analyze and build out."

Jimbo. [added here by Anthere]

Right... that's what was already on Google hosting. Angela 06:33, 29 May 2005 (UTC) Reply

I consider it an update to say that nothing has changed :-) The "update" mean "we have not forgotten, as of "29/05", the situation has not changed :-). Ant

I guess you didn't see this version yesterday. :) Angela 08:12, 29 May 2005 (UTC) Reply

The Amsterdam servers are technically online; they only need configuration.. GerardM 08:42, 29 May 2005 (UTC) Reply

When I said not online, I meant they're not currently in use by the Wikimedia Foundation. Angela 08:52, 29 May 2005 (UTC) Reply

Ping? It's been a year... ~~helix84 17:21, 27 June 2006 (UTC) Reply

Well if Wiki doesn't take Google up on it's offer for free server use, the general public can with Google Page Creator. Page Creator is Google's entry in the web hosting market.