#musicbrainz-devel

/

      • ocharles
        i don't know :) I guess I want people to be able to use it easily, but I'd like changes to come back to me
      • 2012-11-27 33227, 2012

      • warp
        "changes to come back to me" is not enforced by any free software license.
      • 2012-11-27 33241, 2012

      • warp
        ocharles: for now I'd say just pick "GPLv2 or later", because that's what we do for all our musicbrianz stuff.
      • 2012-11-27 33213, 2012

      • ocharles
      • 2012-11-27 33241, 2012

      • warp
        I think the LGPL is almost always a bad idea.
      • 2012-11-27 33250, 2012

      • ocharles
        ok
      • 2012-11-27 33255, 2012

      • warp
        but it depends on what you're trying to achieve with the license.
      • 2012-11-27 33255, 2012

      • ianmcorvidae
        "I want people to be able to use it easily" suggests not-GPL
      • 2012-11-27 33251, 2012

      • warp
        I don't think the GPL is particularly difficult to comply with :)
      • 2012-11-27 33206, 2012

      • ianmcorvidae
        a lot of folks disagree, is all I'm saying :P
      • 2012-11-27 33236, 2012

      • ianmcorvidae
        my take would be: if you want people to be able to use it easily, BSD or something; LGPL as sort of a weird ineffective middle-ground; or GPL if you want some amount of recourse if people don't feel like publishing changes
      • 2012-11-27 33237, 2012

      • ocharles
        it seems I used GPL3 when I created the project, so I'll just stick with that
      • 2012-11-27 33250, 2012

      • warp
        ocharles: hah, ok :)
      • 2012-11-27 33200, 2012

      • ianmcorvidae
        or, since you're not going to actually sue anyone over this, choose at random :P
      • 2012-11-27 33217, 2012

      • warp
        if you go permissive, I would chose Apache 2 over BSD/MIT/X11, because you get the patent protections.
      • 2012-11-27 33221, 2012

      • ianmcorvidae
        (or use BSD or the sqlite license or something appropriately permissive)
      • 2012-11-27 33237, 2012

      • ocharles
        i think I've used MIT in the past
      • 2012-11-27 33246, 2012

      • ocharles
        because a friend did. That's some LOGIC right there
      • 2012-11-27 33252, 2012

      • warp
        :D
      • 2012-11-27 33258, 2012

      • warp
        ocharles: it would help if you at some point at least come to some sort of conclusion about whether you prefer copyleft or permissive.
      • 2012-11-27 33204, 2012

      • ocharles
        yea
      • 2012-11-27 33210, 2012

      • ocharles
        i need to do some reading on that topic
      • 2012-11-27 33250, 2012

      • ocharles writes documentation
      • 2012-11-27 33259, 2012

      • warp writes python
      • 2012-11-27 33201, 2012

      • warp
        and sql
      • 2012-11-27 33251, 2012

      • MBJenkins
        Project musicbrainz-server_beta build #234: FIXED in 35 min: http://ci.musicbrainz.org/job/musicbrainz-server_…
      • 2012-11-27 33252, 2012

      • MBJenkins
        ianmcorvidae: Fix remaining brokenness in tests on hobbes
      • 2012-11-27 33259, 2012

      • warp
        ianmcorvidae: yay!
      • 2012-11-27 33206, 2012

      • hawke joined the channel
      • 2012-11-27 33232, 2012

      • ianmcorvidae
        huh
      • 2012-11-27 33240, 2012

      • ianmcorvidae
        apparently the english wikipedia isn't returning extracts right now
      • 2012-11-27 33243, 2012

      • ianmcorvidae
        for apparently no reason D:
      • 2012-11-27 33202, 2012

      • ianmcorvidae
        (for some pages, seemingly at random D:)
      • 2012-11-27 33237, 2012

      • DremoraLV joined the channel
      • 2012-11-27 33201, 2012

      • MBJenkins
        Project musicbrainz-data build #72: SUCCESS in 46 sec: http://ci.musicbrainz.org/job/musicbrainz-data/72/
      • 2012-11-27 33202, 2012

      • MBJenkins
        ollie: Bump lens dependency upper bound
      • 2012-11-27 33237, 2012

      • CatCat really doesn't like the "read more..." links, those look like the annotation ones and i am expecting the page to just load more text
      • 2012-11-27 33225, 2012

      • ocharles
        on that note, ianmcorvidae any reason you didn't use the same display logic as the annotation?
      • 2012-11-27 33240, 2012

      • ocharles
        actually, I guess the fade out doesn't quite make sense
      • 2012-11-27 33244, 2012

      • ianmcorvidae
        wikipedia pages are huge, and I'm not fetching entire pages in the first place
      • 2012-11-27 33246, 2012

      • ocharles
        cause you don't 'reveal' more of the wiki
      • 2012-11-27 33258, 2012

      • ocharles
        sure, I wasn't suggesting that, I just made the purdy transparent fade
      • 2012-11-27 33205, 2012

      • ocharles
        but then realised this isn't quite the right place
      • 2012-11-27 33205, 2012

      • ianmcorvidae
        ah
      • 2012-11-27 33214, 2012

      • ianmcorvidae
        the extracts are so small I didn't think that would ever be needed
      • 2012-11-27 33230, 2012

      • ianmcorvidae
        unless we wanted to display whole wikipedia pages inline, which I figured we didn't, so :)
      • 2012-11-27 33245, 2012

      • ianmcorvidae
        and yeah, that's a good point on the read more... links, not sure how to improve that
      • 2012-11-27 33203, 2012

      • ianmcorvidae
        there's some bugs with wikipedia's extract endpoint anyway, it appears (see https://beta.musicbrainz.org/artist/b86bf49a-9097… in anything other than German/Spanish, for example -- there's a page, but because of header things on the page it doesn't generate an extract)
      • 2012-11-27 33242, 2012

      • ijabz
        ocharles, regarding licenses I would much prefer something like Apache 2, GPL is pretty horrible IMO because you cannot use anything with GPL unless you make everything that uses it GPL
      • 2012-11-27 33247, 2012

      • ijabz
      • 2012-11-27 33206, 2012

      • warp
        which is exactly why I like it :)
      • 2012-11-27 33234, 2012

      • warp
        although "everything that uses it" is a bit of an exaggeration.
      • 2012-11-27 33210, 2012

      • ijabz
        yeah, but it stops good things happening, for example Apcahe wanted to incorporate my jaudiotagger into an Apcahe project, but even though I was using the less restrictive LGPL that was not acceptable to them
      • 2012-11-27 33222, 2012

      • ianmcorvidae did my first major work in a language where the *LGPL* wasn't sufficently permissible for anyone to actually use it
      • 2012-11-27 33227, 2012

      • ianmcorvidae
        so I guess I have a bias :P
      • 2012-11-27 33257, 2012

      • ijabz
        and because I hadn't written 100% of the code myself and derived some of it from another LGPL project I cannot change the license
      • 2012-11-27 33226, 2012

      • ijabz
        so Apache cannot legally incorporate my library in their open source software
      • 2012-11-27 33226, 2012

      • warp
        ijabz: you cannot easily change the license, no. doesn't mean it's impossible.
      • 2012-11-27 33209, 2012

      • ijabz
        But difficult, if Id used something like APL I wouldn't have this issue
      • 2012-11-27 33231, 2012

      • warp
        ijabz: right, but that is a limitation of apache as an organization. they're the one refusing your work, it's not something inherently wrong with the LGPL.
      • 2012-11-27 33216, 2012

      • warp
        ianmcorvidae: that seems very peculiar.
      • 2012-11-27 33233, 2012

      • warp
        ianmcorvidae: what language was this?
      • 2012-11-27 33237, 2012

      • ianmcorvidae
        Common Lisp
      • 2012-11-27 33203, 2012

      • ianmcorvidae
        http://opensource.franz.com/preamble.html is the preamble that needed to be added to clarify how the LGPL makes any sense for Lisp code
      • 2012-11-27 33241, 2012

      • ijabz
        Im blurring the issue, but if using GPL would be even more of a problem
      • 2012-11-27 33212, 2012

      • warp
        ianmcorvidae: right, but that's mainly because the LGPL is in general difficult to interpret. that would be clearer if you just use the GPL.
      • 2012-11-27 33238, 2012

      • ijabz
        The thing I really don't like about GPL is it makes out it all about sharing code but you can't use that code unless you abide by its restrictive license
      • 2012-11-27 33247, 2012

      • ijabz
        so it isn't really about sharing code at all
      • 2012-11-27 33258, 2012

      • warp
        ijabz: but you understand WHY it is doing that, right?
      • 2012-11-27 33240, 2012

      • ijabz
        Well , sort of, but I don't think it makes much sense, butId be interested in your explanantion
      • 2012-11-27 33233, 2012

      • warp
        I always like the analogy with car mechanics
      • 2012-11-27 33206, 2012

      • warp
        if I buy a car, and something is wrong with it, I should be able to take it to any car machanic and get it fixed. and not be forced to go back to the guy I bought it from.
      • 2012-11-27 33239, 2012

      • warp
        if I buy software, and something is wrong with it, I should be able to take it to any programmer and get it fixed. and not be forced to go back to the guy I bought the software from.
      • 2012-11-27 33208, 2012

      • warp
        the GPL guarantees that the sourcecode is always available to the user of the software, so he/she always has the freedom to make changes to it.
      • 2012-11-27 33243, 2012

      • ijabz
        But this analogy falls down on two counts IMO
      • 2012-11-27 33235, 2012

      • warp waits for ijabz to explain the two counts :)
      • 2012-11-27 33205, 2012

      • ijabz
        1. Whilst it would be good to get the car fixed by any car-mechanic, I dont think it follows that anything in the slightest way related to the car should be published, i.e if I have taken a packet of photographs that happen to be in the car should I make available when and how they were created.
      • 2012-11-27 33252, 2012

      • ijabz
        2. With the car mechanic example you would only make available informations on the parts of the car, you do not prove a car for free. With source code , if freely available you have iftact provided the program itself and therefore a very small chance of getting paid for that work
      • 2012-11-27 33218, 2012

      • warp
        ok, let me answer #1 first.
      • 2012-11-27 33207, 2012

      • warp
        "anything in the slightest way related to the car" is not how the GPL works.
      • 2012-11-27 33217, 2012

      • warp
        the GPL governs a single creative work.
      • 2012-11-27 33207, 2012

      • warp
        if that work is a program, only that program is covered by the GPL. any data you load into it won't be touched by the GPL.
      • 2012-11-27 33224, 2012

      • warp
        any other programs communicating with it through public APIs wouldn't be touched by the GPL.
      • 2012-11-27 33227, 2012

      • warp
        etc..
      • 2012-11-27 33206, 2012

      • ijabz
        But if I use a library with a GPL license as a small part of my application then GPL is enforced on the whole work
      • 2012-11-27 33210, 2012

      • warp
        in the car analogy, a car stereo would be covered by the GPL if the car stereo is connected to the car through some kind of connection unique to that combination.
      • 2012-11-27 33227, 2012

      • ijabz
        Even though I would consider that multiple creative works
      • 2012-11-27 33244, 2012

      • warp
        but if this is an industry standard connection which you can swap out, then they'd still be seperate creative works. and the car radio can have a seperate license from the car.
      • 2012-11-27 33238, 2012

      • ianmcorvidae
        by the terms of the GPL, the... rather standard process of dynamic linking isn't apparently considered industry-standard, then?
      • 2012-11-27 33240, 2012

      • warp
        ijabz: well, yes, if you include a library as part of your program, the resulting application is a single creative work.
      • 2012-11-27 33202, 2012

      • ijabz
        Yes, and I think that is wrong
      • 2012-11-27 33212, 2012

      • ijabz
        whereas I don't mind LGPL
      • 2012-11-27 33232, 2012

      • warp
        ianmcorvidae: how you link is just an implementation detail, it doesn't really matter all that much for deciding whether your application is a derivative of the library or not.
      • 2012-11-27 33257, 2012

      • ianmcorvidae
        warp: according to the GPL, linked in libraries make it the same creative work
      • 2012-11-27 33205, 2012

      • ianmcorvidae
        why is that not an "industry-standard connection" like you're talking about?
      • 2012-11-27 33213, 2012

      • ianmcorvidae
        linking hasn't substantially changed in decades
      • 2012-11-27 33217, 2012

      • hawke_1
        ianmcorvidae: That’s never been tested in court.
      • 2012-11-27 33231, 2012

      • hawke_1
        ianmcorvidae: And it may fail…
      • 2012-11-27 33231, 2012

      • ianmcorvidae
        hawke_1: irrelevant, we're talking philosophy
      • 2012-11-27 33204, 2012

      • ijabz
        Because an app is based on the library, is isn't derived from the library, its just uses the library , I would think of that as one creative works using a number of other creative works
      • 2012-11-27 33248, 2012

      • ijabz
        if the library itself is modified then I agree it is a good thing that the modified library should be freely available
      • 2012-11-27 33219, 2012

      • warp
        ianmcorvidae: I impression is that the FSF presents the GPL and LGPL as being slightly more effective than they really are. lawyers disagree about far the reach of the GPL is in those cases.
      • 2012-11-27 33238, 2012

      • hawke_1
        ianmcorvidae: Then yes, I agree that the GPL is simply incorrect, and linking doesn’t/shouldn’t make the larger work a derivative of the smaller work.
      • 2012-11-27 33240, 2012

      • warp
        s/I/my/
      • 2012-11-27 33255, 2012

      • ianmcorvidae
        as I said to hawke: we're talking philosophy right now; we're talking about "what the FSF presents" pretty immediately :P
      • 2012-11-27 33231, 2012

      • warp
        ijabz: anyway, to answer your #2.
      • 2012-11-27 33200, 2012

      • warp
        ijabz: my personal opinion there is that most users don't read licenses. whether you use the GPL or a restrictive license will have almost no effect on how many copies of the software you sell, and no effect on how many copies of the software are copied.
      • 2012-11-27 33220, 2012

      • warp
        ijabz: but obviously that is just my theory, I doubt that has ever been studied.
      • 2012-11-27 33248, 2012

      • ijabz
        I think you are missing my point, If I use a library licensed with LPGL/APL I don't have to distribute additional source code that I have written
      • 2012-11-27 33252, 2012

      • ijabz
        With GPL I do
      • 2012-11-27 33203, 2012

      • warp
        ianmcorvidae: the idea is that if you're interacting with a standard API, you're at most a derivative of that API -- not of any particular implementation of it.
      • 2012-11-27 33214, 2012

      • warp
        ianmcorvidae: and APIs aren't copyrightable :)
      • 2012-11-27 33220, 2012

      • reosarevok
        I wonder how many people can actually build a relatively complex program from source anyway
      • 2012-11-27 33230, 2012

      • reosarevok
        (about the GPL/source thing)
      • 2012-11-27 33200, 2012

      • reosarevok
        Of course, I guess they could give away own-built packages for free, huh
      • 2012-11-27 33214, 2012

      • hawke_1 wonders what a “standard API” is.
      • 2012-11-27 33215, 2012

      • ijabz
        It doesn't matter whether or not the average user reads the license, but that just one person can take the source and build their own version, original author cut out of the loop
      • 2012-11-27 33229, 2012

      • warp
        ijabz: yes, but distributing with GPL doesn't mean you have to distribute for free. so it has no effect on your "chance of getting paid for that work".
      • 2012-11-27 33200, 2012

      • reosarevok
        warp: admittedly it does if someone else builds the program and puts it for free download in a second site
      • 2012-11-27 33206, 2012

      • reosarevok
        which GPL does allow, right?
      • 2012-11-27 33213, 2012

      • warp
        ijabz: right, but if you sell closed source software, it also just takes one person to take that and put it on the pirate bay.
      • 2012-11-27 33215, 2012

      • ijabz
        Wel it does, because as I say someone else can then build there own version and distribute it for free, so why anyone want the paid version
      • 2012-11-27 33254, 2012

      • dekarl joined the channel
      • 2012-11-27 33255, 2012

      • mb-chat-logger joined the channel
      • 2012-11-27 33219, 2012

      • ijabz
        Yes, but you can put in licensing ecetera if the code are is closed source
      • 2012-11-27 33238, 2012

      • warp
        because people want to get software from a reliable place, where they know they won't get some dodgy virus infected version, where they get updates, where they know their money goes to the author which helps make sure the author can keep working on it.
      • 2012-11-27 33205, 2012

      • reosarevok
        heh
      • 2012-11-27 33211, 2012

      • ijabz
        that is kind of cloud cuckoo land, that is not the real world
      • 2012-11-27 33226, 2012

      • luks
        I'd seen things like picard on thepiratebay, I don't think people care about a reliable source :)
      • 2012-11-27 33239, 2012

      • reosarevok
        Then comes lifehacker, telling people to use jaikoz's trial version to tag a few songs each time instead of paying for it :p
      • 2012-11-27 33250, 2012

      • reosarevok
        (go "money to the author" :p)
      • 2012-11-27 33257, 2012

      • reosarevok
        (it was lifehacker, right?)
      • 2012-11-27 33222, 2012

      • reosarevok
        But yeah, admittedly that's true of closed source too
      • 2012-11-27 33228, 2012

      • warp
        exactly, there are always people out there who will not pay for shit. they don't care about whether they get a GPL'ed version or a pirated copy... they just want it free.
      • 2012-11-27 33231, 2012

      • ijabz
        Yeah sure, but no problem people tried it, and just fixing a few songs is a pain, so some were happy to buy full version
      • 2012-11-27 33240, 2012

      • warp
        it's the people who ARE willing to pay for it you should care about.
      • 2012-11-27 33215, 2012

      • ijabz
        This is just totally unrealistic, I mean loads of people like Picard but it doesn't get that much in donations
      • 2012-11-27 33228, 2012

      • warp
        ijabz: I'm not talking about donations
      • 2012-11-27 33235, 2012

      • warp
        ijabz: I'm talking about selling software.
      • 2012-11-27 33203, 2012

      • luks
        xchat on windows is a good example of paid gpl software failure
      • 2012-11-27 33214, 2012

      • hawke_1
        ijabz: Picard also isn’t presented as for-pay software in the first place.
      • 2012-11-27 33219, 2012

      • reosarevok waves from ychat :p
      • 2012-11-27 33257, 2012

      • hawke_1
        luks: I think that’s because xchat has a model of “it’s free, but fuck you Windows users!”
      • 2012-11-27 33210, 2012

      • warp
        luks: why is it a failure?
      • 2012-11-27 33211, 2012

      • ijabz
        okay, perhaps bad example, but it does have a nag screen, what Im trying to say is most people will not pay for something if there is no real need
      • 2012-11-27 33224, 2012

      • warp
        ijabz: I agree with that.
      • 2012-11-27 33226, 2012

      • luks
        warp: because if you google for "xchat windows", you get silverx as the first link
      • 2012-11-27 33231, 2012

      • luks
        which is a free build for windows
      • 2012-11-27 33244, 2012

      • warp
        luks: I understand that, but I don't see how that makes xchat for windows a failure.
      • 2012-11-27 33201, 2012

      • ijabz
        so if you make it very easy for someone to provide same software with no charge, they are not going to pay