On Sat, May 7, 2011 at 2:44 PM, Ben Sturmfels wrote:

On 02/05/11 15:34, rdbrown@pacific.net.au wrote:

...

Diomidis D. Spinellis's latest IEEE Software column "Choosing and Using Open Source Components" is up at his blog http://www.spinellis.gr/blog/20110501 and may interest.

Thanks for pointing this out Rodney. Hope we see you at the upcoming discussion group!

My apologies for the question: haven't had enough time 'til now to attend any of the meetings. Can you please update me with date/time/location for the next meeting?

It's a well written article, but does seem to be a little freedom-agnostic. I also worry about Spinellis' slightly twisted interpretation of the GPL:

Others (licenses), like the GNU licenses, play well with other software licensed as open source but make life difficult for proprietary offerings. This is especially true if you want to distribute your work to others as a shrink-wrapped package, such as Microsoft Office, or as an embedded software product, like a set-top box. In such cases the only GNU-licensed components you can easily use are unmodified dynamically linked libraries licensed under the so-called GNU Lesser General Public License (LGPL). You get considerable more leeway with GNU-licensed software if you don’t distribute a product but instead offer a service (like Google) or simply use your system privately within your organization.

The GNU GPL says that you can't incorporate GPL licensed code into proprietary programs.

Slight correction here: yes, you can/may, as long as you don't distribute the result of incorporation in any kind. And I'd argue that this is *not* outside the spirit of the free software. I'd argue that running/providing a GPL-ed software *as a service* and asking money in return is still in the spirit of free software and, to some extent, beneficial for the free software that is used (exposure) and/or for the "consumers" of such a service. Examples: heaps of hosting providers offering LAMP (on quite low prices) - are they operating outside the spirit of free software? Are they even "hurting" the spirit of free software? Are there any problems with non-free additions *specific* to the exploitation environ? E.g. for a hosting service provider, some specific additions to capture the usage and, based on the captured data, to interface with their specific billing systems. Must these additions be published as well? No matter the re-usability considerations or risks for the users/consumers? My point is: where do you draw the line of what is inside or outside of the spirit of free software for a non-free/closed-source part added to a GPL-ed software, offered as a service and never distributed? Finally, I'm more worried about the following in the blog: <quote>Although it’s tempting, try to avoid modifying the open source code to fit your needs; you don’t want to end up maintaining another large component on your own.</quote> Now, this *IS* outside the spirit of free software (at least the way I see it). I can see the angle Spinellis is coming, but I do have huge issues with the form he expressed it. It is one thing to say: "If you develop your own customizations, you face the risk of broken compatibility with future releases of the free software" and a different thing to say: "Stay out of customizations! Believe me, you don't want to spend anything in making the software better *even for you* much less for anyone else". The first is a fair warning. The second is bordering FUD of the same sort as Microsoft's "get the facts right" campaign.

Cheers, Ben

Adrian

On Mon, May 9, 2011 at 5:43 PM, Ben Sturmfels wrote:

On 09/05/11 11:20, Adrian Colomitchi wrote:

...

On Sat, May 7, 2011 at 2:44 PM, Ben Sturmfels wrote:

On 02/05/11 15:34, rdbrown@pacific.net.au wrote:

...

Diomidis D. Spinellis's latest IEEE Software column "Choosing and Using

...

Open Source Components" is up at his blog http://www.spinellis.gr/blog/20110501 and may interest.

Thanks for pointing this out Rodney. Hope we see you at the upcoming discussion group!

My apologies for the question: haven't had enough time 'til now to attend any of the meetings. Can you please update me with date/time/location for the next meeting?

Thurs 19 May, 6:30pm at State Library. Look forward to meeting you! :)

For dates beyond that: http://www.softwarefreedom.com.au/free-software-melb/

It's a well written article, but does seem to be a little

...

...

freedom-agnostic. I also worry about Spinellis' slightly twisted interpretation of the GPL:

Others (licenses), like the GNU licenses, play well with other software licensed as open source but make life difficult for proprietary offerings. This is especially true if you want to distribute your work to others as a shrink-wrapped package, such as Microsoft Office, or as an embedded software product, like a set-top box. In such cases the only GNU-licensed components you can easily use are unmodified dynamically linked libraries licensed under the so-called GNU Lesser General Public License (LGPL). You get considerable more leeway with GNU-licensed software if you don’t distribute a product but instead offer a service (like Google) or simply use your system privately within your organization.

The GNU GPL says that you can't incorporate GPL licensed code into proprietary programs.

Slight correction here: yes, you can/may, as long as you don't distribute the result of incorporation in any kind. And I'd argue that this is *not* outside the spirit of the free software.

The difference here is just terminology. For me, a program can't be proprietary until it has been distributed. So I think we agree there.

With this choice of terminology, total agreement. This settles the matter.

I'd argue that running/providing a GPL-ed software *as a service* and

...

asking money in return is still in the spirit of free software and, to some extent, beneficial for the free software that is used (exposure) and/or for the "consumers" of such a service. Examples: heaps of hosting providers offering LAMP (on quite low prices) - are they operating outside the spirit of free software? Are they even "hurting" the spirit of free software?

I encourage people to provide network services using free software. I do it myself in my business. I object to turning free software into a network service specifically to get around the requirements of the GPL (as the article suggests).

[...]

Perhaps networks services would be an interesting topic for the next discussion group.

Yes, it would be interesting. You see, I can't believe that "turning a GPL-ed software into exclusively a service *just to get around the GPL*" would be indeed a risk to worry about: most of the time, the "pain" of operating a consistent service at a non-trivial level of usage/security exceeds the effort to develop/maintain the software (sort of saying: the cost of developing particular customization is far exceeded by the cost of operating the customized software... ). But I have a hunch that there are other risks (for the free software) that comes with "service-alization"

Finally, I'm more worried about the following in the blog:

...

<quote>Although it’s tempting, try to avoid modifying the open source code to fit your needs; you don’t want to end up maintaining another large component on your own.</quote> Now, this *IS* outside the spirit of free software (at least the way I see it). I can see the angle Spinellis is coming, but I do have huge issues with the form he expressed it. It is one thing to say: "If you develop your own customizations, you face the risk of broken compatibility with future releases of the free software" and a different thing to say: "Stay out of customizations! Believe me, you don't want to spend anything in making the software better *even for you* much less for anyone else". The first is a fair warning. The second is bordering FUD of the same sort as Microsoft's "get the facts right" campaign.

That's a good point. There is a trade-off between maintaining customisations and having to use software that doesn't fit you. Customising is the best solution in some situations. Perhaps Spinellis' wording was a little off.

Customising may be an important way a free software evolve to cover a broader set of needs (as long as the same needs are shared by a larger user base): discouraging effort in this direction *is* contrary to the spirit of free software (sort of saying that, in my personal opinion, the wording is not a bit but way off).

Ben

Adrian