Hi,
In the past, when an account was expired, it was still possible to access its public_html. This has now changed. Any HTTP requests to an expired account will return an error page indicating that the account has expired.
The files in public_html are not deleted, and will become accessible again if the account is ever un-expired.
- river.
Hi,
River Tarnell a écrit :
In the past, when an account was expired, it was still possible to access its public_html. This has now changed. Any HTTP requests to an expired account will return an error page indicating that the account has expired.
The "gallery" tool integrated into WM Commons is now unavailable, I assume it is a direct consequence of this change. http://toolserver.org/~daniel/WikiSense/Gallery.php?wikifam=commons.wikimedia.org&img_user_text=Eusebius
Is there a way to make it work again, by reactivating the host account or moving the tool on someone else's account (and modifying the relevant config on Commons)? I don't see it as a vital tool, but I'm sure many users (and admins) find it useful.
Regards, Eusebius
River Tarnell wrote:
Eusebius:
The "gallery" tool integrated into WM Commons is now unavailable, I assume it is a direct consequence of this change.
Yes, but unintentional. This should be fixed now.
CheckUsage, at least, doesn't work. http://toolserver.org/~daniel/WikiSense/CheckUsage.php
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 10-02-04 08:46 AM, River Tarnell wrote:
In the past, when an account was expired, it was still possible to access its public_html. This has now changed. Any HTTP requests to an expired account will return an error page indicating that the account has expired.
The files in public_html are not deleted, and will become accessible again if the account is ever un-expired.
In the past, when people disappeared, their tools would (maybe) keep working. Now that we *know* they will stop working once the account is expired, do we know how requests for sourcecode for a tool will be handled, in case someone wants to revive it?
And this is another reason multi-maintainer accounts are superior! Use them!
- -Mike
Mike.lifeguard:
In the past, when people disappeared, their tools would (maybe) keep working. Now that we *know* they will stop working once the account is expired, do we know how requests for sourcecode for a tool will be handled, in case someone wants to revive it?
Currently we don't have a good policy for this. Many (most?) tools have no copyright header or license, which makes it difficult to distribute them. There have been some proposals in the past, like assuming tools without licenses will use a particular license, but I'm not too fond of that.
The solution I like best is to require that all tools be MMTs, but I imagine some users would object to that ;-)
This is probably something that WM-DE needs to create a policy for.
- river.
One solution would be to make free licensing a prerequisite to obtaining a toolserver account. And then make it mandatory to make all source-code world readable (minus config files containing passwords of course).
Currently we don't have a good policy for this. Many (most?) tools have no copyright header or license, which makes it difficult to distribute them.
On 4 February 2010 17:14, Daniel Schwen lists@schwen.de wrote:
One solution would be to make free licensing a prerequisite to obtaining a toolserver account. And then make it mandatory to make all source-code world readable (minus config files containing passwords of course).
In my opinion, this is unreasonable.
Firstly, let me say that I support free/open source software and use it heavily.
However, I feel that toolserver users should retain the *freedom* to choose how to license projects. Perhaps the docs and so on ought to be adjusted to emphasise the importance and value of free licensing, but there should be no automatic compulsion by dint of ToS to give certain freedoms to one's work. While I, certainly, would open source any further Wikimedia related tools I were to write, I strongly feel this should be a matter of personal preference.
Also, if a user "vanishes" and their tool becomes unmaintained, the correct course of action unless the code is freely licensed and has been released to the public is, I think, to accept that the tool is gone and write a new one (as a MMT!). Situations where tools become unusable for various reasons highlight the importance of open-sourcing code, but whether or not to license code in such a way should remain a *free choice*.
A possibility, while retaining the freedom to license one's code how one wishes, is that the toolserver could exercise its freedom to tell people "no" when they request accounts, or to close accounts. To do so purely on the basis of the licensing of the code - as long as there isn't evidence that licences upon other code are being breached - strikes me as counter-productive, since tools which are for whatever reason closed-source, could still be immensely useful to the community.
Thanks
Martin
On Thu, Feb 4, 2010 at 9:38 PM, Martin Peeks martinp23@googlemail.com wrote:
However, I feel that toolserver users should retain the *freedom* to choose how to license projects. Perhaps the docs and so on ought to be adjusted to emphasise the importance and value of free licensing, but there should be no automatic compulsion by dint of ToS to give certain freedoms to one's work. While I, certainly, would open source any further Wikimedia related tools I were to write, I strongly feel this should be a matter of personal preference.
At a minimum, though, this should be opt-out, just for pragmatic reasons. The problem isn't that TS users don't want to freely license their tools, it's that they vanish and we don't have the opportunity to ask them. If we said "All your tools are GPLv3 or later, unless you specifically say otherwise", then most tools would end up being freely-licensed, and when someone disappears, someone else could take over the tool.
I wouldn't have any problem at all with saying that if you want to write closed-source software related to Wikimedia, you can do it on your own dime and not using Wikimedia-DE's hardware, software, or administration budget. But it's not my decision to make, of course.
On 5 February 2010 02:45, Aryeh Gregor Simetrical+wikilist@gmail.com wrote:
On Thu, Feb 4, 2010 at 9:38 PM, Martin Peeks martinp23@googlemail.com wrote:
However, I feel that toolserver users should retain the *freedom* to choose how to license projects. Perhaps the docs and so on ought to be adjusted to emphasise the importance and value of free licensing, but there should be no automatic compulsion by dint of ToS to give certain freedoms to one's work. While I, certainly, would open source any further Wikimedia related tools I were to write, I strongly feel this should be a matter of personal preference.
At a minimum, though, this should be opt-out, just for pragmatic reasons. The problem isn't that TS users don't want to freely license their tools, it's that they vanish and we don't have the opportunity to ask them. If we said "All your tools are GPLv3 or later, unless you specifically say otherwise", then most tools would end up being freely-licensed, and when someone disappears, someone else could take over the tool.
The default copyright stance, unless a licence specifies otherwise, is "All Rights Reserved". I don't think we have the right to enforce a licence that is all about freedom unless a user opts-in.
I wouldn't have any problem at all with saying that if you want to write closed-source software related to Wikimedia, you can do it on your own dime and not using Wikimedia-DE's hardware, software, or administration budget. But it's not my decision to make, of course.
Closed source software can be as good as open source software - do remember that. And closed source software doesn't have to be commercial. While (imo) WM-DE should support free and open source software, this can be done in other ways. For example, by using a free and open source webserver rather than the current Zeus (which is to my knowledge closed-source - at least I cannot find source on their site).
To me, this is all about freedoms - that's all.
Martin
Well I would think that we would be able to have them "opt-in" to the free licence in the same way that wikimedia does (you agree to it when you signup/submit something). We would just be giving them the opportunity to opt-out of that and still use the service. I would tend to agree with having an opt-out function, mostly for legitimate concerns like Pathoschilds.
James
James Alexander james.alexander@rochester.edu jamesofur@gmail.com 100 gmail invites and no one to give them to :( let me know if you want one :)
On Thu, Feb 4, 2010 at 9:53 PM, Martin Peeks martinp23@googlemail.com wrote:
On 5 February 2010 02:45, Aryeh Gregor Simetrical+wikilist@gmail.com wrote:
On Thu, Feb 4, 2010 at 9:38 PM, Martin Peeks martinp23@googlemail.com wrote:
However, I feel that toolserver users should retain the *freedom* to choose how to license projects. Perhaps the docs and so on ought to be adjusted to emphasise the importance and value of free licensing, but there should be no automatic compulsion by dint of ToS to give certain freedoms to one's work. While I, certainly, would open source any further Wikimedia related tools I were to write, I strongly feel this should be a matter of personal preference.
At a minimum, though, this should be opt-out, just for pragmatic reasons. The problem isn't that TS users don't want to freely license their tools, it's that they vanish and we don't have the opportunity to ask them. If we said "All your tools are GPLv3 or later, unless you specifically say otherwise", then most tools would end up being freely-licensed, and when someone disappears, someone else could take over the tool.
The default copyright stance, unless a licence specifies otherwise, is "All Rights Reserved". I don't think we have the right to enforce a licence that is all about freedom unless a user opts-in.
I wouldn't have any problem at all with saying that if you want to write closed-source software related to Wikimedia, you can do it on your own dime and not using Wikimedia-DE's hardware, software, or administration budget. But it's not my decision to make, of course.
Closed source software can be as good as open source software - do remember that. And closed source software doesn't have to be commercial. While (imo) WM-DE should support free and open source software, this can be done in other ways. For example, by using a free and open source webserver rather than the current Zeus (which is to my knowledge closed-source - at least I cannot find source on their site).
To me, this is all about freedoms - that's all.
Martin
Toolserver-l mailing list (Toolserver-l@lists.wikimedia.org) https://lists.wikimedia.org/mailman/listinfo/toolserver-l Posting guidelines for this list: https://wiki.toolserver.org/view/Mailing_list_etiquette
On 5 February 2010 02:57, James Alexander jamesofur@gmail.com wrote:
Well I would think that we would be able to have them "opt-in" to the free licence in the same way that wikimedia does (you agree to it when you signup/submit something). We would just be giving them the opportunity to opt-out of that and still use the service. I would tend to agree with having an opt-out function, mostly for legitimate concerns like Pathoschilds.
You hint upon another obvious counter-argument: namely that the Wikimedia Projects demand CC licensing on text and they get on well. Obviously the situation isn't the same here - namely that CC licensing is a necessity for the goal of the project.
I would submit that F/OS licensing is not a *necessity* for tools. It's very nice though, and I agree with the obvious benefits.
Aryeh, earlier: "if you want to write closed-source software related to Wikimedia, you can do it on your own dime and not using Wikimedia-DE's hardware, software, or administration budget"
The cost of administrating an opt-out programme and enforcing free/open-source rules would be rather onerous on an administration budget, I feel.
Again, I think that education as to the benefits of open sourcing code to precipitate a complete change in attitude in those who habitually go closed-source is far better than forcing it upon them.
Thanks Martin
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256
Martin Peeks wrote:
The cost of administrating an opt-out programme and enforcing free/open-source rules would be rather onerous on an administration budget, I feel.
Not really, it's easy to tell what's being used or not from access logs, see if link to source is provided anywhere and if it's there, ok, if not, deluser. Takes all of two minutes. Make it easier by requiring that tools provide a link to their source available somewhere online.
Hi,
On 5 February 2010 03:08, Q overlordq@gmail.com wrote:
Not really, it's easy to tell what's being used or not from access logs, see if link to source is provided anywhere and if it's there, ok, if not, deluser. Takes all of two minutes. Make it easier by requiring that tools provide a link to their source available somewhere online.
If you are being serious by "if not, deluser", then I worry. Such a response goes contrary to the whole idea of keeping tools going.
Is the proposal to: (a) Force users to open-source software under their own choice of licence. (b) Assume (supported by ToS) that all code is under a given licence. (c) Care not abut the licensing but decree that all code must be world-readable (d) Other!
Thanks Martin
Martin Peeks wrote:
Hi,
On 5 February 2010 03:08, Q overlordq@gmail.com wrote:
Not really, it's easy to tell what's being used or not from access logs, see if link to source is provided anywhere and if it's there, ok, if not, deluser. Takes all of two minutes. Make it easier by requiring that tools provide a link to their source available somewhere online.
If you are being serious by "if not, deluser", then I worry. Such a response goes contrary to the whole idea of keeping tools going.
Is the proposal to: (a) Force users to open-source software under their own choice of licence. (b) Assume (supported by ToS) that all code is under a given licence. (c) Care not abut the licensing but decree that all code must be world-readable (d) Other!
Thanks Martin
I still dont see why you feel it's morally wrong to say that "If you want to use our servers, it's required to use an open source license so that if get ran over, have a heart attack, or just get tired of running it, other people can take over your code so that the community can continue to rely on it."
to agree with having an opt-out function, mostly for legitimate concerns like Pathoschilds.
Pathoshilds concern is not really valid. 1) per Mike.lifeguards comment, 2) we would not have to actively disseminate the code. The issue is sharing between toolserver users. Potentially harmful tools like pathoschild's (which seem to bee against TS rules anyways) cannot be kept from being distributed just through license terms in any case. Either you trust your fellow TS users, or if not, do not bring this kind of software to the toolserver in the first place.
On Thu, Feb 4, 2010 at 10:05 PM, Daniel Schwen lists@schwen.de wrote:
Pathoshilds concern is not really valid. 1) per Mike.lifeguards comment, 2) we would not have to actively disseminate the code. The issue is sharing between toolserver users. Potentially harmful tools like pathoschild's (which seem to bee against TS rules anyways) cannot be kept from being distributed just through license terms in any case.
Synchbot does not violate the terms of service in any way. It neither requests nor collects user credentials; users provide temporary access directly to me, knowing the risks of doing so. The Toolserver is simply the platform for providing this service, which you can read about at < http://meta.wikimedia.org/wiki/User:Pathoschild/Scripts/Synchbot
. (Account creation is an optional extra service.)
The license indicates what freedoms the author intends for his tools. It makes no sense to license a tool for redistribution if you don't want it redistributed. Trust between Toolserver users is irrelevant here, since home directories are not world-readable.
On Thu, Feb 04, 2010 at 11:13:58PM -0500, Jesse (Pathoschild) wrote:
On Thu, Feb 4, 2010 at 10:05 PM, Daniel Schwen lists@schwen.de wrote:
Pathoshilds concern is not really valid. 1) per Mike.lifeguards comment, 2) we would not have to actively disseminate the code. The issue is sharing between toolserver users. Potentially harmful tools like pathoschild's (which seem to bee against TS rules anyways) cannot be kept from being distributed just through license terms in any case.
Synchbot does not violate the terms of service in any way. It neither requests nor collects user credentials; users provide temporary access directly to me, knowing the risks of doing so. The Toolserver is simply the platform for providing this service, which you can read about at < http://meta.wikimedia.org/wiki/User:Pathoschild/Scripts/Synchbot
. (Account creation is an optional extra service.)
The license indicates what freedoms the author intends for his tools. It makes no sense to license a tool for redistribution if you don't want it redistributed. Trust between Toolserver users is irrelevant here, since home directories are not world-readable.
Up to the lawyers to make a license that people *can* use, to give the rights to the toolserver administrators (aka WM-DE) to use/edit their tools?
That can then be one of the accepted licenses for toolserver tools.
Regards,
Andre
The default copyright stance, unless a licence specifies otherwise, is "All Rights Reserved". I don't think we have the right to enforce a licence that is all about freedom unless a user opts-in.
Of course we have. When a user obtains an account on the toolserver he effectively enters contract with Wikimedia Deutschland. Of course licensing can be tied to that. How can you suggest otherwise?!
Closed source software can be as good as open source software - do remember that. And closed source software doesn't have to be commercial. While (imo) WM-DE should support free and open source
This is _completely_ besides the point. Please do not make this a heise-forum-style idiology debate. There are purely pragmatic reasons for free licensing here: 1) avoiding dying tools when maintainers leave. 2) fostering synergies through code sharing
Of course you can take the stance to say "just write the tool again". What a waste of time! Developer resources should be values higher than this.
On 5 February 2010 02:59, Daniel Schwen lists@schwen.de wrote:
The default copyright stance, unless a licence specifies otherwise, is "All Rights Reserved". I don't think we have the right to enforce a licence that is all about freedom unless a user opts-in.
Of course we have. When a user obtains an account on the toolserver he effectively enters contract with Wikimedia Deutschland. Of course licensing can be tied to that. How can you suggest otherwise?!
I mean a moral right.
Closed source software can be as good as open source software - do remember that. And closed source software doesn't have to be commercial. While (imo) WM-DE should support free and open source
This is _completely_ besides the point. Please do not make this a heise-forum-style idiology debate. There are purely pragmatic reasons for free licensing here: 1) avoiding dying tools when maintainers leave. 2) fostering synergies through code sharing
Of course you can take the stance to say "just write the tool again". What a waste of time! Developer resources should be values higher than this.
You seem to have missed the point that I agree with the benefits of open-source and am a strong supporter of it and hence am well acquainted with the benefits.
Forcing it upon people is bad. Re-writing a tool is onerous and can be avoided by *educating* users to open source their software, and perhaps gently prodding them from time-to-time to do so.
Martin
Martin Peeks:
The default copyright stance, unless a licence specifies otherwise, is "All Rights Reserved". I don't think we have the right to enforce a licence that is all about freedom unless a user opts-in.
We can require that all users use a free license for their tools, make this clear to them at account request time, and if they fail to do so, they are simply breaking the rules, which we already have a policy to deal with. (And no, it's not as simple as "just userdel".)
Closed source software can be as good as open source software - do remember that.
The issue here is nothing to do with whether the software is good or not. No one is suggesting that tools will somehow become much better because they're open source, and (unless I missed it) no one is suggesting that tools should be open source for ideological reasons, only for pragmatic reasons.
The issue is that when users leave, and their tools have no free license, those tools go away and cannot be revived. So, we need to balance freedom for users with the ability to transfer ownership of tools after the owner leaves.
With that in mind, there are three options: * Do nothing. * Require all tools to be open source * Require all tools to be multi-maintainer projects, and further require that MMTs must be open source.
If we do anything other than "do nothing", the downside is that all users who are unwilling to open source their tools will leave. While that's not a problem in itself, it means any such tools might no longer be available (if the owner can't find other hosting), or might only work in a reduced form without database access. This is clearly detrimental to Wikimedia projects, so it's something we want to avoid.
The downside of staying with "do nothing" is that all non-open-source tools will no longer be available once the owner leaves. This is an active and ongoing problem that leaves many useful tools inaccessible or broken.
Now, while I haven't done any actual research on this, it seems quite likely to me that the number of tools we would lose from requiring open source is much lower than the number we currently lose from inactive maintainers. Given that we can't have both options, it's clear that requiring open source is the better one.
The actual details of how to do this is a separate discussion and not relevant to this issue... but fear not, I will definitely not support a policy that requires a viral license like the GPL for tools.
While (imo) WM-DE should support free and open source software, this can be done in other ways. For example, by using a free and open source webserver rather than the current Zeus (which is to my knowledge closed-source - at least I cannot find source on their site).
It seems unlikely to me that Zeus will go out of business in the near future. On the other hand, users go "out of business" all the time, leaving their widely-used, unlicensed tools broken and unmaintainable. So, leaving aside ideological reasons (which I have no time for), there is very little reason to change from ZWS, but a very good reason to require freely licensed tools.
- river.
On 5 February 2010 12:02, River Tarnell river.tarnell@wikimedia.de wrote:
We can require that all users use a free license for their tools, make this clear to them at account request time, and if they fail to do so, they are simply breaking the rules, which we already have a policy to deal with.
Note that it is also possible some tools _cannot_ be freely licensed, because they contain/use third-party code/libraries not under a free license. This might be a purely theoretical concept (no, I do not known about anything specific, and all my tools might be freely-licensed), but possibly worth considering, esp. if a hard policy is being prepared.
-- [[cs:User:Mormegil | Petr Kadlec]]
Petr Kadlec:
Note that it is also possible some tools _cannot_ be freely licensed, because they contain/use third-party code/libraries not under a free license.
This is only the case if they actually derive the third-party code. Just linking to it is fine (possibly not for GPL, but for more reasonable licenses). (There are also open source licenses that allow licensed code to be mixed with non-free code in the same executable, as long as it's not in the same file.)
Even so, if the tool is using the library in the first place, it should be possible to construct a license that allows distribution and modification, even if it's not a fully free license. The only time this isn't possible is if the code were licensed to only a single person, which I imagine should be quite unlikely.
(Example: a tool using code that only allows non-commercial use cannot be open source, but it can still be licensed in a way that allows it to be shared amongst Toolserver and other users.)
- river.
On 5 February 2010 11:02, River Tarnell river.tarnell@wikimedia.de wrote:
Martin Peeks:
Closed source software can be as good as open source software - do remember that.
The issue here is nothing to do with whether the software is good or not. No one is suggesting that tools will somehow become much better because they're open source, and (unless I missed it) no one is suggesting that tools should be open source for ideological reasons, only for pragmatic reasons.
Actually, "we" have had a very long-term rule that we don't use non-OSS software as a part of the Wikimedia "stack". This rule is so old it pre-dates the Foundation (I remember it being a concern that the new-fangled Foundation wouldn't necessarily take it into account in future). It has always been an ideologically-based rule, even if it has also had practical needs.
I know that a lot of people will claim that the Toolserver's tools aren't part of the software used as a part of the Wikimedia wikis. I disagree. To the end-user, the software that we the community use to link from geo-tagged articles in Wikipedia, or find categories, or other things. Sure, lots of these are editor- rather than reader-focussed, but I don't think that there is an easy line to draw.
If people want a shell account to run "cool" tools that happen to do Wikimedia-related things, there's no particular need for them to have TS access. There are lots of free or cheap shell providers out there. TS access is a privilege, and we should expect those who have the advantages of this privilege to work within the same rules that we expect from other members of the community.
But then, I've never written code for the toolserver, running code on my own shells when I've needed it. So maybe I'm biased. :-)
Yours,
On 05.02.2010, 18:39 James wrote:
Actually, "we" have had a very long-term rule that we don't use non-OSS software as a part of the Wikimedia "stack". This rule is so old it pre-dates the Foundation (I remember it being a concern that the new-fangled Foundation wouldn't necessarily take it into account in future). It has always been an ideologically-based rule, even if it has also had practical needs.
I know that a lot of people will claim that the Toolserver's tools aren't part of the software used as a part of the Wikimedia wikis. I disagree. To the end-user, the software that we the community use to link from geo-tagged articles in Wikipedia, or find categories, or other things. Sure, lots of these are editor- rather than reader-focussed, but I don't think that there is an easy line to draw.
If people want a shell account to run "cool" tools that happen to do Wikimedia-related things, there's no particular need for them to have TS access. There are lots of free or cheap shell providers out there. TS access is a privilege, and we should expect those who have the advantages of this privilege to work within the same rules that we expect from other members of the community.
But then, I've never written code for the toolserver, running code on my own shells when I've needed it. So maybe I'm biased. :-)
Yours,
As the matter of fact, both Toolserver AND Wikimedia cluster already use non-free software. Using an ideologically correct technology is a good thing, but some also believe that using a functional technology is good too.
James Forrester:
Actually, "we" have had a very long-term rule that we don't use non-OSS software as a part of the Wikimedia "stack".
Actually, if such a rule exists, it was and still is frequently broken. The response I got when I asked about this was not "we don't use closed source software", but "we use open source software where possible".
Furthermore, it does not apply to the Toolserver; I specifically confirmed this with WM-DE when we started. Nonetheless, let's assume that it does.
If we have a tool which requires Toolserver access to work, and also uses non-open-source code which prevents it from being open source itself, there are two options: we can allow the tool to run while not being open source, or we can deny access and the tool will not exist.
If we choose the first option, we gain a useful tool, and have lost nothing.
If we choose the second option, we gain "ideological compliance", but lose a useful tool.
If you think the second option is better, I don't really know how to respond. This is the sort of open-source pyrrhic victory that the GNU people are so fond of. If your mission is to promote open source software and destroy proprietary software, perhaps it makes sense. That is not our mission.
If people want a shell account to run "cool" tools that happen to do Wikimedia-related things, there's no particular need for them to have TS access. There are lots of free or cheap shell providers out there.
Actually, a lot of tools *do* need Toolserver access. That's why it exists. (If something could be equally easily hosted somewhere else, we consider this a point against granting a new account.)
TS access is a privilege, and we should expect those who have the advantages of this privilege to work within the same rules that we expect from other members of the community.
I see conflict between this statement and the Toolserver allowing non-open-source tools where necessary.
- river.
On Thu, Feb 4, 2010 at 9:53 PM, Martin Peeks martinp23@googlemail.com wrote:
The default copyright stance, unless a licence specifies otherwise, is "All Rights Reserved". I don't think we have the right to enforce a licence that is all about freedom unless a user opts-in.
We have the right to require anything we like of people who use our hardware. WMDE is under no obligation, moral or otherwise IMO, to permit people to use their resources in ways they don't think are best for the toolserver.
On Fri, Feb 5, 2010 at 10:39 AM, James Forrester james@jdforrester.org wrote:
Actually, "we" have had a very long-term rule that we don't use non-OSS software as a part of the Wikimedia "stack".
This rule is not applicable to the toolserver, which indeed runs mostly on proprietary software. I personally wish it were applicable, but it's not and never has been.
On Fri, Feb 5, 2010 at 9:49 AM, DaB. WP@daniel.baur4.info wrote:
If somebody is interested, the DRAFT can be found under [1]. I updated it today and added some stuff from the discussion here. I'm of corse interested in response (and fixing of my bad english), but I can't guarantee that it will ever accept by the WMDE. . . . [1] https://wiki.toolserver.org/view/Toollizenz/draft/en
"The source code of every tool is licensed under a free (like BSD) or GPL-compatible[1] license by default"
I don't think this is clear, or legally enforceable. I'm pretty sure you have to give a specific license, e.g., "GPLv3 or later", or "MIT license". You can't agree to license your code under an unspecified free license.
"Source code is exempted from this if it is explicitly licensed. A tool's license can be explicitly licensed in a comment at the top of its source code, or visibly in its user interface."
Maybe you should also allow users to create ~/LICENSE or something that gives a default license for all their tools.
Also, something I'm not clear on:
"To keep with the successful spirit of free licensing for all Wikimedia projects, the tools on the Toolserver are also freely licensed. . . . Source code is exempted from this if it is explicitly licensed."
The second sentence suggests that the code can be explicitly licensed as "Proprietary, all rights reserved", while the first sentence suggests the opposite.
code, but whether or not to license code in such a way should remain a *free choice*.
Getting a TS account or not getting a TS account already is a free choice. It is not unreasonable at all to tie certain policies to the account. As a matter of fact this is already being done. [1]
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 10-02-04 12:24 PM, River Tarnell wrote:
Many (most?) tools have no copyright header or license, which makes it difficult to distribute them. There have been some proposals in the past, like assuming tools without licenses will use a particular license, but I'm not too fond of that.
We could require that users freely license their code - then you would know at least that the four freedoms apply and redistributing it wouldn't be so hairy. But of course, one is supposed to know precisely /which/ free license is used...
- -Mike
On Thu, Feb 4, 2010 at 7:46 PM, Mike.lifeguard mike.lifeguard@gmail.com wrote:
But of course, one is supposed to know precisely /which/ free license is used...
Choose the most restrictive (AGPL3 for example). Then users themselves can also choose to release them under a more free license.
There might be some legal caveats though. Can you force users to release their work under a certain license by requiring that in the TOS? Can you change the TOS for existing users to force them to release new works under a certain license? And can you force users to retroactively license previously created works by changing that in the TOS? I would presume that 1) and 2) are possible, but I doubt 3). Ianal though.
Bryan
On Thu, Feb 4, 2010 at 1:46 PM, Mike.lifeguard mike.lifeguard@gmail.com wrote:
We could require that users freely license their code - then you would know at least that the four freedoms apply and redistributing it wouldn't be so hairy.
I don't think this is a good idea. Most of my tools are freely licensed and open-source, as you can see in the footer at http://toolserver.org/~pathoschild/stalktoy/, but there are a few tools I do not want to publish the source code for. Synchbot is a good example; it registers an account on every wiki in two minutes (faster with changes), which I wouldn't want in the hands of our crosswiki attack username vandals.
It's fine to recommend free licensing and open source, but we should have the option not to for such cases.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 10-02-04 03:07 PM, Jesse (Pathoschild) wrote:
Synchbot registers an account on every wiki in two minutes (faster with changes), which I wouldn't want in the hands of our crosswiki attack username vandals.
I didn't think you ran that from the toolserver, since that bot requires the user's login data.
Regardless, keep in mind that even if the code is freely licensed, there is no requirement for you to distribute it. Simply don't publish the code.
But now we get into questions of how to know which code should be published and which shouldn't be if the author is gone... Do we simply assume it can be redistributed safely unless there's a note in the header? I have a headache, I leave the pondering on this to others
- -Mike
Well ... imho, the only safe way to handle such issues is to check whether there is an open sourced license clause in the source code. Only then you can safely use code after an account is expired.
I consider all my work as published under GPLv3, although not everywhere I included the license clause (I should definitely make work of it).
Regards Annabel
----- Original Message ---- From: Mike.lifeguard mike.lifeguard@gmail.com To: toolserver-l@lists.wikimedia.org Sent: Thu, February 4, 2010 8:14:09 PM Subject: Re: [Toolserver-l] Changes to expired accounts web hosting
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 10-02-04 03:07 PM, Jesse (Pathoschild) wrote:
Synchbot registers an account on every wiki in two minutes (faster with changes), which I wouldn't want in the hands of our crosswiki attack username vandals.
I didn't think you ran that from the toolserver, since that bot requires the user's login data.
Regardless, keep in mind that even if the code is freely licensed, there is no requirement for you to distribute it. Simply don't publish the code.
But now we get into questions of how to know which code should be published and which shouldn't be if the author is gone... Do we simply assume it can be redistributed safely unless there's a note in the header? I have a headache, I leave the pondering on this to others
- -Mike
_______________________________________________ Toolserver-l mailing list (Toolserver-l@lists.wikimedia.org) https://lists.wikimedia.org/mailman/listinfo/toolserver-l Posting guidelines for this list: https://wiki.toolserver.org/view/Mailing_list_etiquette
Дана Thursday 04 February 2010 20:14:09 Mike.lifeguard написа:
But now we get into questions of how to know which code should be published and which shouldn't be if the author is gone... Do we simply assume it can be redistributed safely unless there's a note in the header? I have a headache, I leave the pondering on this to others
I see a difference between publicly accessible and inaccessible code. Free license might be required for publicly accessible code, and other code might remain fully copyrighted.
Hello At Thursday 04 February 2010 21:29:52 DaB. wrote:
Mike.lifeguard:
In the past, when people disappeared, their tools would (maybe) keep working. Now that we know they will stop working once the account is expired, do we know how requests for sourcecode for a tool will be handled, in case someone wants to revive it?
Currently we don't have a good policy for this. Many (most?) tools have no copyright header or license, which makes it difficult to distribute them. There have been some proposals in the past, like assuming tools without licenses will use a particular license, but I'm not too fond of that.
The solution I like best is to require that all tools be MMTs, but I imagine some users would object to that ;-)
This is probably something that WM-DE needs to create a policy for.
I requested such a policy (with a draft (even a english one)) more then 6 months ago. Maybe now if some important tools will vanish, WM-DE will start to think about it.
Added Tim who is responsible for technical-questions.
Sincerly, DaB.
Hello, At Friday 05 February 2010 15:45:21 DaB. wrote:
I requested such a policy (with a draft (even a english one))
If somebody is interested, the DRAFT can be found under [1]. I updated it today and added such stuff from the discussion here. I'm of corse interested in response (and fixing of my bad english), but I can't guarantee that it will ever accept by the WMDE.
Sincerly, DaB.
River Tarnell wrote:
Hi,
In the past, when an account was expired, it was still possible to access its public_html. This has now changed. Any HTTP requests to an expired account will return an error page indicating that the account has expired.
The files in public_html are not deleted, and will become accessible again if the account is ever un-expired.
- river.
It doesn't seem to have changed. gmaxwell account expired time ago. However, its files can still be accessed http://toolserver.org/~gmaxwell/election_analysis/2008/GRAPH_3_totals.png
Not that i'm complaining about that since I am still using http://toolserver.org/~gmaxwell/cgi-bin/deletedimage.py despite it's only half-working... Any takers?
*Gregory Maxwell hasn't "disappeared" so don't be afraid just for licensing issues. ;)
Platonides:
It doesn't seem to have changed. gmaxwell account expired time ago. However, its files can still be accessed http://toolserver.org/~gmaxwell/election_analysis/2008/GRAPH_3_totals.png
# acctexp gmaxwell The account "gmaxwell" will expire on Tuesday, 01 June 2010.
- river.
River Tarnell wrote:
Platonides:
It doesn't seem to have changed. gmaxwell account expired time ago. However, its files can still be accessed http://toolserver.org/~gmaxwell/election_analysis/2008/GRAPH_3_totals.png
# acctexp gmaxwell The account "gmaxwell" will expire on Tuesday, 01 June 2010.
- river.
He must have reenabled it. *goes to bug him*
toolserver-l@lists.wikimedia.org