Menu
▾
▴
clucene-developers
|
From: Maxim K. <ma...@fl...> - 2003-07-03 01:51:30
|
Hi guys, Sorry I caused a bit of trouble with the cryptic question. What I meant is that clucene being under lgpl is not a good result for the proprietary project I am working on. I also did not make any legal claims that clucene should be under apache like lucene, all I meant is that it would make sense to go along with the original authors' intentions. So, are we good with Apache? Max. |
|
From: Max K. <ma...@to...> - 2003-07-03 14:29:25
|
Hi guys, since I asked the original q-n and caused some confusion let me add more: when I said 'proprietary project' I meant the project I am working on not clucene. Having to confirm to LGPL will be very restrictive. And when I said that clucene should be under apache I did not make any legal claim (I am not a lawer :) I was only saying that this would make sense according to the original authors' intention. Doug? What do you think? thanks, max. |
|
From: Doug C. <cu...@lu...> - 2003-07-03 17:37:18
|
Max Khesin wrote: > Doug? What do you think? I hate dealing licensing issues, but they are important. This is Ben's decision, as he's written the clucene code. Some tips: - One should be careful about who copyright is assigned to, as only that party can change the copyright and license. So once you assign the copyright to the Apache Foundation, or the FSF, you cannot change it. I couldn't find any copyright statements in the CLucene code. It's probably a good idea to add one to each file. - If you use GPL or LGPL then folks (like Max) will complain endlessly. I initially selected GPL for Lucene without thinking much about it. Folks complained and I switched to LGPL. They complained less, but still complained. With the move to Apache the complaints stopped. So, if your interest, like mine, is in developing code that folks will just use, then Apache will save you some hassle dealing with compaliners. But, depending on your free software political agenda, this may not be acceptable to you. Doug |
|
From: Ben v. K. <be...@vi...> - 2003-07-03 23:07:38
|
OK, I think i'm beginning to understand the issues behind licensing opensource... what a hassle. To me, flexibility and unhindered use of clucene is important. In my understanding, that is what LGPL is (unhindered use) - (though it seems that i am wrong!). I too started off on GPL, as that was the "suggested" license on sourceforge. I now understand that GPL is too restrictive, but why change to Apache? So Doug, it seems like i went through the same rigmarole as you. Selecting GPL without to much thought, going to LGPL, now to Apache... I have no political agenda when it comes to clucene, so if you (doug) think that is what people want then i see no harm in it. Ben ----- Original Message ----- From: Doug Cutting <cu...@lu...> To: <clu...@li...> Sent: Friday, July 04, 2003 3:37 AM Subject: Re: [CLucene-dev] Re: license (clarification) > Max Khesin wrote: > > Doug? What do you think? > > I hate dealing licensing issues, but they are important. > > This is Ben's decision, as he's written the clucene code. > > Some tips: > - One should be careful about who copyright is assigned to, as only > that party can change the copyright and license. So once you assign the > copyright to the Apache Foundation, or the FSF, you cannot change it. I > couldn't find any copyright statements in the CLucene code. It's > probably a good idea to add one to each file. > - If you use GPL or LGPL then folks (like Max) will complain > endlessly. I initially selected GPL for Lucene without thinking much > about it. Folks complained and I switched to LGPL. They complained > less, but still complained. With the move to Apache the complaints > stopped. So, if your interest, like mine, is in developing code that > folks will just use, then Apache will save you some hassle dealing with > compaliners. But, depending on your free software political agenda, > this may not be acceptable to you. > > Doug > > > > ------------------------------------------------------- > This SF.Net email sponsored by: Free pre-built ASP.NET sites including > Data Reports, E-commerce, Portals, and Forums are available now. > Download today and enter to win an XBOX or Visual Studio .NET. > http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01 > _______________________________________________ > CLucene-developers mailing list > CLu...@li... > https://lists.sourceforge.net/lists/listinfo/clucene-developers > > |
|
From: <rg...@sd...> - 2003-07-05 01:53:39
|
Hi,
I just joined the mailing list, however I have read all the
messages in the archives.
I had a few comments about licenses; but I want to preface
them by saying that the clucene developers can expect me to
work on clucene regardless of whether it is under the GPL,
LGPL, BSD or Apache licenses.
First, I am unsure of why Max Khesin thinks the LGPL is
incompatible with a proprietary product in a way that the
Apache license isn't. Both licenses would require that you do
something when distributing a proprietary program that used
CLucene. The Apache requires that notices appear in the
proprietary software, the LGPL requires the receiver of the
proprietary software be able to get the source code to
CLucene. Personally, I think placing as zip file of CLucene
source on a CD or web page is less burdensome than cluttering
up my "About" sceen. Would the names "Apache" and "Apache
Software Foundation" be replaced with Ben van Klinkehn ?
Some licenses allow the switching later on to others, such as
the latest BSD licensed software can be re-licensed as GPL (or
practically anything), and LGPL can be re-licensed as GPL, but
since Apache requires that notice it can't be switched to GPL
or LGPL later.
You can offer the software under all three choices, at the
user's discretion. This is called "dual licensing" or
"multiple licensing."
Someone made reference to the original writers of Lucene and
their intentions. When a piece of software code is re-written
completely, so that you change every single line, which is
what you do in switching languages, it is a new copyrightable
work not a derived work of the previous. (This is a grand
simplification, of course, and there are surely exceptions.)
An example is the Berkeley group that took AT&T Unix and
re-worked it until it was the separate work of BSD Unix.
Sometimes people are confused about this issue and think they
can't own code they have written while having someone else's
example in front of them, because they have heard discussions
of "clean room programming" and "virgin programmers" (virgin
programmers are so inexperienced they are known never to have
studied common examples). The issues of clean room and virgin
refer to trade secrets and whether reverse engineering was
used, not copyrights. We should be grateful to the Apache
group for saving us from all that by making their work Free
Software.
In any case, the original Lucene and the Apache project is
focused on code meant to be run on a server, and most ordinary
users of the code don't have it on their computers; while
CLucene, being a simple library not in Java, is well suited to
being used in applications distributed everywhere. Users of
software via a server don't have much freedom to modify the
code anyway, and the people running the server are more likely
to be in a corporation, so the two environments naturally lead
to different outlooks on licenses.
This message is long because I wanted to get all my license
opinions into one message. My future emails will be about
using and modifying CLucene :)
--Rob
|
|
From: Max K. <ma...@to...> - 2003-07-06 01:37:46
|
Hi Rob, I am not a legalist, but after looking at lgpl it appears that any source modifications have to be published, too. That's the part that is not good for a proprietary project. Apache has no such restictions. On 4 Jul 2003, Rob Ristroph wrote: > > Hi, > > I just joined the mailing list, however I have read all the > messages in the archives. > > I had a few comments about licenses; but I want to preface > them by saying that the clucene developers can expect me to > work on clucene regardless of whether it is under the GPL, > LGPL, BSD or Apache licenses. > > First, I am unsure of why Max Khesin thinks the LGPL is > incompatible with a proprietary product in a way that the > Apache license isn't. Both licenses would require that you do > something when distributing a proprietary program that used > CLucene. The Apache requires that notices appear in the > proprietary software, the LGPL requires the receiver of the > proprietary software be able to get the source code to > CLucene. Personally, I think placing as zip file of CLucene > source on a CD or web page is less burdensome than cluttering > up my "About" sceen. Would the names "Apache" and "Apache > Software Foundation" be replaced with Ben van Klinkehn ? > > Some licenses allow the switching later on to others, such as > the latest BSD licensed software can be re-licensed as GPL (or > practically anything), and LGPL can be re-licensed as GPL, but > since Apache requires that notice it can't be switched to GPL > or LGPL later. > > You can offer the software under all three choices, at the > user's discretion. This is called "dual licensing" or > "multiple licensing." > > Someone made reference to the original writers of Lucene and > their intentions. When a piece of software code is re-written > completely, so that you change every single line, which is > what you do in switching languages, it is a new copyrightable > work not a derived work of the previous. (This is a grand > simplification, of course, and there are surely exceptions.) > An example is the Berkeley group that took AT&T Unix and > re-worked it until it was the separate work of BSD Unix. > Sometimes people are confused about this issue and think they > can't own code they have written while having someone else's > example in front of them, because they have heard discussions > of "clean room programming" and "virgin programmers" (virgin > programmers are so inexperienced they are known never to have > studied common examples). The issues of clean room and virgin > refer to trade secrets and whether reverse engineering was > used, not copyrights. We should be grateful to the Apache > group for saving us from all that by making their work Free > Software. > > In any case, the original Lucene and the Apache project is > focused on code meant to be run on a server, and most ordinary > users of the code don't have it on their computers; while > CLucene, being a simple library not in Java, is well suited to > being used in applications distributed everywhere. Users of > software via a server don't have much freedom to modify the > code anyway, and the people running the server are more likely > to be in a corporation, so the two environments naturally lead > to different outlooks on licenses. > > This message is long because I wanted to get all my license > opinions into one message. My future emails will be about > using and modifying CLucene :) > > --Rob > > > ------------------------------------------------------- > This SF.Net email sponsored by: Free pre-built ASP.NET sites including > Data Reports, E-commerce, Portals, and Forums are available now. > Download today and enter to win an XBOX or Visual Studio .NET. > http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01 > _______________________________________________ > CLucene-developers mailing list > CLu...@li... > https://lists.sourceforge.net/lists/listinfo/clucene-developers > |
|
From: <rg...@sd...> - 2003-07-06 02:28:12
|
>>>>> "Max" == Max Khesin <ma...@to...> writes: Max> Max> Hi Rob, Max> I am not a legalist, but after looking at lgpl it appears that Max> any source modifications have to be published, too. That's the Max> part that is not good for a proprietary project. Apache has no Max> such restictions. Here is how I plan to go about selling a proprietary software that uses CLucene as a library: 1) The "free trial" or "download version" of the software will have a link to the source forge CLucene site somewhere on the page. Not hidden or in fine print, but somewhere below the "Download Now" button there will be a paragraph that says something like "This software uses a Free Software library called CLucene, and many of it's features would not be available without that project." It has to say the version number. ( If I have changes that aren't in the official tree, then I have to host the tar file of source myself, which I regard as mild encouragement to keep the project from splitting, but not too burdensome if I have to do it. ) 2) Any registered version obtained by download will have a similar link. 3) Any version sold or given away via CDROM will have the tar file of source on the CD, and a README that tells you were it came from and that the latest version is on sourceforge. Basically, any method you use to obtain my software, the CLucene source will be available via the same method. That should be enough to satisfy the LGPL. There is also something about a written offer to provide source code, I can add that to any README files if all my documentation is electronic and add it in the legalese at the end of the printed manual if there is one. --Rob P.S. If you want to get better advice than I can give on these issues, I recommend the gnu.misc.discuss newsgroup. |
|
From: Max K. <ma...@to...> - 2003-07-06 04:21:09
|
Hi Rob, all of this is great if you are not modifying the original code. Let's say you are tightly integrating your code with clucene, via source rather than a shared library. You may be ok with lgpl, or somebody can theoretically come and demand you to publish all your code under 'derived works' clause. Maybe you will win the case, after shelling out your hard-earned money for lawers or maybe you will loose or give up. This is very improbable, but you cannot have real peace of mind without a short clear license like apache. And apache would not prevent people like me from contributing/improving the main open-source apache. This is what I would LIKE to do. Without having a license which will give me a peace of mind I would rather, frankly, start from scratch. Anyway, like Doug said, 'with lgpl people complained and with apache the complains stopped'. So I am going to have to keep complaining :). BTW a commonly used dataset for searching can be found here: http://www.daviddlewis.com/resources/testcollections/reuters21578/ |
|
From: <rg...@sd...> - 2003-07-06 05:41:29
|
>>>>> "Max" == Max Khesin <ma...@to...> writes: Max> Max> Hi Rob, Max> all of this is great if you are not modifying the original Max> code. Let's say you are tightly integrating your code with Max> clucene, via source rather than a shared library. You may be ok Max> with lgpl, or somebody can theoretically come and demand you to Max> publish all your code under 'derived works' clause. Maybe you Max> will win the case, after shelling out your hard-earned money for Max> lawers or maybe you will loose or give up. This is very Max> improbable, but you cannot have real peace of mind without a Max> short clear license like apache. There's no "maybe" about it, if you distribute a binary that is produced partitially from LGPL code and doesn't provide all the source, you will be in violation of the license, unless incorporates the LGPL code through the specific method of linking the library (and provided source to that library, including any changes you made). It would not be a "somebody" who sued, the only person who could sue you for that would be the author or authors of CLucene; so at least you know who to buy off ;) Max> And apache would not prevent people like me from Max> contributing/improving the main open-source apache. Both licenses would allow you to share your changes if you wanted, so that's not really the issue is it ? The issue is that you want to NOT share your changes, and the LGPL doesn't allow that and the Apache does. It's not my code and I'm plunging ahead with either license. However, if you wanted some of my code under the Apache license, I would make you pay for it -- my reasoning being that I get something for my code, either access to better versions of it, or money. Max> This is what I would LIKE to do. Without having a license which Max> will give me a peace of mind I would rather, frankly, start from Max> scratch. Anyway, like Doug said, 'with lgpl people complained and Max> with apache the complains stopped'. So I am going to have to keep Max> complaining :). Swish-e is also debating going through a license change discussion, you might go over to there forum and complain there too :) They have mentioned the BSD license, which would statisfy your requirements. I have conducted a pretty energetic search of all the various index / search tools looking for what was under which license. The Managing Gigabytes stuff is GPL, but there are a small number of authors so you might ask about a license change there. The original SWISH was "GPL or LGPL" according to the announcement: http://www.rice.edu/sw/swish/patches/kevinh.19970916.html And htdig and just about everyone else is GPL, and when they aren't GPL they are written in Java or Perl or another inappropriate language. Or else they are designed to be used through a web interface, and would need modification to make them into a general library. There are small projects for searching specific things like bibliographies, but none of them seem to have the robustness and development that lucene/clucene have. Max> BTW a commonly used dataset for searching can be found here: Max> Max> http://www.daviddlewis.com/resources/testcollections/reuters21578/ I had heard of this, I think it was referenced in the very good book "Managing Gigabytes". ( http://www.mds.rmit.edu.au/mg/ ) I was wanting something easy for sanity testing, and something I knew I could re-distribute to you guys, so I could make tar file and send it and say "see, search for this and then grep for it, something's wrong." It's actually only 8 MB, that's not bad maybe I'll just use it. --Rob |
|
From: Max K. <ma...@to...> - 2003-07-06 06:11:03
|
> There's no "maybe" about it, if you distribute a binary that is > know who to buy off ;) If a commericial license was available I would consider it, although at this early stage my funds are very limited. I just do not think the way things are now clucene is ready to offer a commercial license, but Ben may disagree :). > Both licenses would allow you to share your changes if you wanted, so > that's not really the issue is it ? The issue is that you want to NOT > share your changes, and the LGPL doesn't allow that and the Apache > does. The problem is that some changes are not really an improvement of the open source package as its goals are defined. e.g if I stick clucene in some cool documentation software I am working on my prog will be subject to either the open-source req. of lgpl or to the inconvenience of using a linked binary. > It's not my code and I'm plunging ahead with either license. However, > if you wanted some of my code under the Apache license, I would make > you pay for it -- my reasoning being that I get something for my code, > either access to better versions of it, or money. I do not see how you can pull this off under apache. apache means it's open-source. and even if you do not post your code but sell it to me under apache i can just go ahead and post it myself. what I have seen some people do is offer the code under either GPL OR a commercial license. This way people who do not want to be burdened by GPL can just go ahead and by the GPL-free license. > Swish-e is also debating going through a license change discussion, > you might go over to there forum and complain there too :) They have > mentioned the BSD license, which would statisfy your requirements. Maybe I will :). Is it written in an 'appropriate' language? > and just about everyone else is GPL, and when they aren't GPL they are > written in Java or Perl or another inappropriate language. :). Yes, these are inappropriate. (the authors of original lucene may disagree) > I had heard of this, I think it was referenced in the very good book > "Managing Gigabytes". ( http://www.mds.rmit.edu.au/mg/ ) I have both 1st and recently got 2d editions. My 'dayjob' is working for a company which developed the 1st (and still the best :) commercial implementation of JBIG2 image compression. Many principles of symbolic compression being covered in MG, we are longtime friends :) > It's actually only 8 MB, that's not bad maybe I'll just use it. The advantage is you can say things like 'twice as fast as the X search engine on the Reuters corpus', similar to the 'CCITT 8' image set when it comes to compression (static huffman codes in fax machines were optimized based on these 8 images). cheers, max. |
|
From: <rg...@sd...> - 2003-07-06 16:33:51
|
>>>>> "Max" == Max Khesin <ma...@to...> writes: Max> Max> The problem is that some changes are not really an improvement of Max> the open source package as its goals are defined. e.g if I stick Max> clucene in some cool documentation software I am working on my Max> prog will be subject to either the open-source req. of lgpl or to Max> the inconvenience of using a linked binary. It all comes down to what Title 17 refers to as a "derived work." If CLucene were just a separate executable called via VB script in your product's documentation window, no problems. If you integrate it too far it is a derived work. The FSF makes the claim that all dynamic linking is deriving a new work from the library, but the LGPL explicitly gives up the author's right to works _derived by linking_ (but not by other means). And it is possible that at least some uses of dynamic linking aren't derived works inspite of the FSF's position. It isn't clear where the line is, not because of the license, but because Title 17 purposely doesn't make it explicit. >> However, if you wanted some of my code under the Apache license, I >> would make you pay for it -- my reasoning being that I get >> something for my code, either access to better versions of it, or >> money. Max> Max> I do not see how you can pull this off under apache. apache means Max> it's open-source. and even if you do not post your code but sell Max> it to me under apache i can just go ahead and post it Max> myself. Yes, that is what I was suggesting. You can offer it under the GPL and wait until someone asks about a different license, then charge money. If you give it to them under the Apache license you will probably only sell it once, but sometimes a single large commercial license fee is all you can get anyway. Sometimes they try to sell a traditional commercial license in which the licensee can't re-distribute the source. Max> what I have seen some people do is offer the code under Max> either GPL OR a commercial license. This way people who do not Max> want to be burdened by GPL can just go ahead and by the GPL-free Max> license. Exactly. Doesn't MySQL do this ? >> Swish-e is also debating going through a license change discussion, >> you might go over to there forum and complain there too :) They have >> mentioned the BSD license, which would statisfy your requirements. Max> Maybe I will :). Is it written in an 'appropriate' language? C. I think it is technically less advanced than MG or Lucene, because it is older. It does have a stemmer though. I don't know the size of the index relative to the body of work. http://swish-e.org/ (Look at the link at the bottom of the page for the License discussion) I think I'm sticking with CLucene though, the LGPL is fine for me and I think it is more advanced than Swish-e. Max> I have both 1st and recently got 2d editions. My 'dayjob' is Max> working for a company which developed the 1st (and still the best Max> :) commercial implementation of JBIG2 image compression. Many Max> principles of symbolic compression being covered in MG, we are Max> longtime friends :) Have you tried to use the software from the site ? I have found the "get scripts" which you need to index things to be buggy and brittle. >> It's actually only 8 MB, that's not bad maybe I'll just use it. Max> Max> The advantage is you can say things like 'twice as fast as the X Max> search engine on the Reuters corpus', similar to the 'CCITT 8' Max> image set when it comes to compression (static huffman codes in Max> fax machines were optimized based on these 8 images). I'm downloading it. --Rob |
Thanks for helping keep SourceForge clean.
X