Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Posted Sep 2, 2016 11:14 UTC (Fri) by nix (subscriber, #2304)In reply to: Contemplating the possible retirement of Apache OpenOffice by nim-nim
Parent article: Contemplating the possible retirement of Apache OpenOffice
You should never buy small-fish innovative software after a few years of Oracle, IBM, etc ownership. It will invariably have degenerated in a feature-packed slow-as-a-snail clusterfuck that requires expensive consultants to deploy and never quite manages to hit the promised target.I think this is unfair. What's happening here isn't some sort of Large Corporation Evil scattered over a project by virtue of its ownership. What's happening here is more in the nature of different sorts of management. In my case, I've seen a project with only a few dozen staff in a fairly small company utterly drowned in box-ticking featuritis of the sort you mentioned; I've also seen projects fly along under corporate ownership (the team was small but the company employed fifty thousand people).
The key, I think, is the nature of the management chain and more generally the number of people who choose to put their oar in on changes. If that is small and the chain is shallow, even a really big company can do stuff very fast, almost like a startup. If the thing is stakeholders up the wazoo, most of whom necessarily don't actually understand the project or are making project decisions for unrelated reasons, you have sclerosis. It's just that big companies are more prone to have long management chains and horrible politics than small ones -- but, again, that is not necessarily the case, and small parts of large companies can retain the go-getting dammit-full-speed-ahead spirit, often by explicit upper-management fiat or simply because nobody else in the company really cares much about what they're doing or knows they're there.
But Apache... Apache's processes are more or less designed to get input from everyone come what may, and thus produce sclerosis. Letting everyone have input seems like a really praiseworthy thing, but it has a cost, and given enough of it the project becomes immobile. (My uncle worked for an insurer once on a business-critical project where, because it was so critical, every change, even changes with no functional effect, had to be signed off by committees numbering 40+ people. Unsurprisingly the project was more or less frozen in aspic. Forget development: even trivial maintenance to fix terrible problems became impossible.)
Posted Sep 2, 2016 11:40 UTC (Fri)
by nim-nim (subscriber, #34454)
[Link]
And big corporations can be quite affective and agile on some projects (but that's the exception not the rule and usually only happens on projects linked to the company core competency).
However big generalistic software behemots like IBM and Oracle combine size with diversification, they do not have a clearly defined corporate focus (except database so Oracle). That's a lethal combination for any third-party software they become the owner of.
Posted Sep 2, 2016 12:27 UTC (Fri)
by jimjag (guest, #84477)
[Link] (69 responses)
Posted Sep 2, 2016 13:44 UTC (Fri)
by davidgerard (guest, #100304)
[Link] (65 responses)
Posted Sep 2, 2016 14:14 UTC (Fri)
by jimjag (guest, #84477)
[Link] (64 responses)
In any case, the idea behind AOO was to create a permissively licensed OO implementation that could be used and consumed by everyone; a common, shared core that people could contribute back to, in true FOSS fashion (especially for people who like weak-copyleft) and allowing them to focus on their own implementations, w/o bothering with, or worrying about, the common bits.
Now, as we know, that was the goal: having a central place where the common bits could be shared and worked on, while allowing for alternative end-user implementations to grow and thrive, as they focus on their specific end-user audience. What happened is that, due to bad blood based on how people were treated by Oracle, the misconception that Apache "stole" OO from TDF, and self-serving trolls who personally benefited from creating division between the communities, this sharing never happened. In particular, despite being pro weak-copyleft, LO refused to do willingly what they force others to do: give back their improvements and patches to the "upstream" code. As such, instead of being a shared resource, AOO was forced into going it alone, which, of course, created further division.
Posted Sep 2, 2016 14:31 UTC (Fri)
by pizza (subscriber, #46)
[Link]
You leave out that "upstream" AOO was licensed differently to LO, making "giving back" considerably more problematic due to license incompatibilities.
Posted Sep 2, 2016 19:40 UTC (Fri)
by gerv (guest, #3376)
[Link] (60 responses)
Gerv
Posted Sep 2, 2016 21:39 UTC (Fri)
by jimjag (guest, #84477)
[Link] (56 responses)
Permissive licensed code is always universal donor.
Posted Sep 2, 2016 22:45 UTC (Fri)
by davidgerard (guest, #100304)
[Link] (3 responses)
Posted Sep 2, 2016 23:58 UTC (Fri)
by jimjag (guest, #84477)
[Link] (2 responses)
FOR THE RECORD: I am NOT saying that Noel Grandin is saying this or doing this!
Posted Sep 3, 2016 19:00 UTC (Sat)
by dtardon (subscriber, #53317)
[Link]
So suddenly it's not enough if a LibreOffice developer provides his patches under ASL, but he's also required to bring them to AOO on a silver plate?
Btw, LibreOffice developers have been in the same situation re AOO commits. And they have coped with it.
Posted Sep 3, 2016 22:45 UTC (Sat)
by davidgerard (guest, #100304)
[Link]
Posted Sep 3, 2016 0:07 UTC (Sat)
by Wol (subscriber, #4433)
[Link] (51 responses)
Add to that, LO is a renamed Go-OO, which was MPL-licenced from the beginning. So there was never a question of "LO chose a licence that was incompatible with AOO" - the fact is LO was there *long* before, and it was AOO that chose the incompatible licence ... (although, being Apache, it wasn't the Apache foundation's choice - the fact remains however that it was the people behind OO who chose to be incompatible ...)
Cheers,
Posted Sep 3, 2016 2:24 UTC (Sat)
by jimjag (guest, #84477)
[Link] (50 responses)
Posted Sep 3, 2016 10:08 UTC (Sat)
by Wol (subscriber, #4433)
[Link] (49 responses)
Apache had the advantage (a massive advantage) of actually OWNING THE COPYRIGHT in OO. *THAT* is the point you seem wilfully blind to. LO and the TDF *COULD* *NOT* relicence Go-OO/LO without a massive administrative headache that any number of developers could simply have refused to participate in, and block the process.
Simply put, Apache could re-licence OO by managerial fiat. TDF had no hope in hell of relicencing Go-OO/LO without a *massive* administrative headache, and with no guarantee of success.
(The only reason the relicencing of LO from LGPL to MPL was able to succeed, was because once AOO relicenced OO to Apache, it was just a matter of an audit to make sure all the code really was either OO and compatible with the MPL, or LO and licenced under the MPL. A completely different kettle of fish from trying to get loads of developers, quite a few of which are probably dead :-(, to agree to change their licence.)
Yes, they could have asked for all NEW contributions to be licenced Apache2, but that would still have given the AOO people a massive headache because every cherry-pick into AOO would have been needed to be checked for Apache compatibility. And I think by that time there was too much bad blood :-(
Cheers,
Posted Sep 3, 2016 13:30 UTC (Sat)
by dtardon (subscriber, #53317)
[Link] (48 responses)
Posted Sep 3, 2016 14:18 UTC (Sat)
by jimjag (guest, #84477)
[Link] (47 responses)
At that point, if they had any desire to contribute back to AOO, they would have had the opportunity to ask that developers allow their *patches* and code to be triple licensed. As noted on the above page, their firm resistance to permissive licensing did not allow for that, although they took full advantage of that permissive license to do what they wanted. Again, the ALv2 does not force contributions back, but the hope and intent is that people who consume ALv2 code will be altruistic enough to do so. We see that in the case of LO, this did not happen.
Posted Sep 3, 2016 19:14 UTC (Sat)
by dtardon (subscriber, #53317)
[Link] (20 responses)
Sorry, but who's "they"? TDF? AFAIK TDF has never prohibited contributing to AOO. And IMHO it's task of AOO, not TDF, to propagate AOO among developers...
> As noted on the above page, their firm resistance to permissive licensing did not allow for that, although they took full advantage of that permissive license to do what they wanted.
The work to bring usable AOO commits to LibreOffice--and it has not been an easy work, despite what you hint at--has been done by individual developers, not by some amorphous "they".
Posted Sep 3, 2016 20:09 UTC (Sat)
by shmget (guest, #58347)
[Link] (1 responses)
https://cgit.freedesktop.org/libreoffice/core/log/?h=aoo/...
YTD for 2016 there is 166 commits on the aoo/trunck branch. 7 of them made it to lo -- 4 of them by Damjan Jovanovic who did commit some interesting and useful fixes (thanks) -- with an annotation of 'merged as: <sha>' the sha being the sha of that equivalent commit in master
the rest is either marked as
To put it in perspective: YTD for 2016 there has been 10516 commit on master.. so 7/10516 = 0.066%, so much for the 'upstream' myth.
Posted Sep 4, 2016 3:21 UTC (Sun)
by bronson (subscriber, #4806)
[Link]
Posted Sep 3, 2016 20:33 UTC (Sat)
by Wol (subscriber, #4433)
[Link] (17 responses)
> Sorry, but who's "they"? TDF? AFAIK TDF has never prohibited contributing to AOO. And IMHO it's task of AOO, not TDF, to propagate AOO among developers...
JIm's point - a very valid one imho - is that the board of TDF could have asked LO contributors to triple-licence - Apache/MPL/LGPL.
The problem is, Jim seems to be (like all the "best" generals) wilfully ignorant of the realities on the ground. Many LO devs were upset with Sun/Oracle because of their imperviousness to contributions. Many LO devs were upset because they felt that giving OO to the Apache foundation was a kick in the teeth for them for all the work they'd done on Go-OO. And then we have the general trolling of LO by people at Apache.
And if you read this, Jim, you need to stop being evasive, and start taking personal responsibility instead of coming over as a corporate PR spokesdroid. Back in this article, you were asked a DIRECT question about what YOU thought. So, in reply, you pointed to a mail server and said "you'll find the answer there". Leaving the reader to search for a needle in a haystack! Most people here have a very poor opinion of Apache and AOO. You're not helping. You come over as a nice guy pushing the corporate line. Not likely to make you many friends here.
And I am quite happy to say that, in my own experience, all the trolling was pretty much one way certainly on LWN. The LO guys mostly ignored AOO, the AOO "spokes-troll" delighted in making a nuisance of himself.
Plus, the AOO devs have done nothing to actively help LO. Even if the LO devs were willing to triple-licence (and I expect a lot of them aren't, now), it would be up to the AOO devs to cherry-pick from the LO code base. So they'd have to check that the contributor had triple-licenced. They would have to have checked that the commit didn't depend on a Go-OO commit. They would have had to have checked that the commit didn't depend on a dev who had refused to relicence ...
If AOO had opened by asking the LO people to triple licence, and had asked the LO people to help with the grunt work of converting OOO into AOO, and hadn't invested so much effort in actively infuriating the LO devs, then things might have got along much better.
The problem is, AOO started with a codebase that was laden with technical debt, and they actively harassed a project that had invested a lot of effort into getting rid of that debt from the same original codebase. If they'd asked the LO people to help out to mutual benefit, they would probably have had a friendly reception. Instead, they asked the LO people to throw away all the hard work they'd done cleaning up the codebase, and to start the SAME work again from scratch. And when the LO guys (unsurprisingly) refused, they got all upset and obnoxious.
Cheers,
Posted Sep 3, 2016 21:29 UTC (Sat)
by spaetz (guest, #32870)
[Link]
Thing is that if I am a proponent of (weak) copyleft licensing, I will *not* triple license under the Apache - a permissive - license which effectively nullifies my copyleft provisions. In this case we could simply ditch the rest and just use the Apache license in the first place.
So, a contributor valueing copyleft principles would of course lose out in the proposed scheme. So it's nothing personal against AOO and conspiracy theories and misguided. As LO contributors often explicitely value copyleft, I believe this is one of the reasons why so little ends up Apache-licensed. I know it is the case for me.
Posted Sep 3, 2016 22:26 UTC (Sat)
by jimjag (guest, #84477)
[Link] (15 responses)
The question you refer to appears to be "Please explain your view of how AOO ended up here." Since "here" is clear, I will address here as in "why OO ended up at Apache" as well as here as in "this state".
1. Why at Apache.
Again, the full discussions on whether or not to accept OO into Apache are open and public. All this happened in June of 2011. There were hundreds of posts about it, which can be read at https://lists.apache.org/list.html?general@incubator.apac... .
Esp look at:
o https://lists.apache.org/thread.html/c2d942f99e00e6ca97a8...
So, in general, the reason why OO ended up at Apache is that there was a need (and still is, BTW) for a permissively licensed office suite that could serve as a core, commonly shared Office implementation that could be used, leveraged and consumed by the entire OO eco-system. Instead of people creating their own one-off, AOO could serve as a base core that people could build upon. As such, it could serve as a central sharing place for code, again to benefit the entire OO community. If the desire was to create as many FOSS alternatives for MSO, then having such a permissively licensed base was key. Even the FSF admits that if wide adoption and free/open standards are important, more permissively licensed implementations are better (https://www.gnu.org/licenses/license-recommendations.en.html). Finally, it was hoped that Apache could help "heal the wounds" as it were between the OO and LO communities.
No let's be clear: Oracle was not going to donate OO to TDF/LO, for a number of reasons. One of which is that TDF didn't legally exist at the time. Another is that Oracle wanted to have the OO codebase under a permissive license. As far as I know there were talks between Oracle and TDF, but to no avail. Knowing that it is pretty clear that if Apache had not accepted OO, Oracle would likely have simply shuttered it or sold it to IBM, neither of which benefits the OO community.
Which brings us to...
2. Why at this state.
As was mentioned, it was hoped that, as such, there would be great cooperation between TDF and AOO. That did not happen. There was way too much bad blood between LO/TDF and Oracle which got transferred over to Apache. FUD was spread about secret deals between IBM, Oracle and Apache, to discourage potential developers from hacking AOO and instead going to LO. The license-wars (copyleft vs. permissive) were played to great glory portraying AOO as enemies of FOSS and LO being loyal to the cause. This caused a handful of AOO developers and aficionados to go just-as-postal and start trolling LO in return. Various people with money in the game fanned the flames to ensure their "investments" on both sides paid out.
All in all, the good will, the spirit and hope of co-operation and coop-itition never happened.
Finally, some big committers to AOO got frustrated that they were contributing useful stuff to AOO which then got pretty immediately consumed by LO, when, at the same time, there was resistance (ranging from minor to extensive) from stuff going the other way. Feeling that TDF wasn't "playing fair" these committers stopped, leaving a vacuum in AOO.
Posted Sep 3, 2016 23:22 UTC (Sat)
by DOT (subscriber, #58786)
[Link]
Whatever comes next, I hope OpenOffice users won't be left with unsupported software that won't ever be upgraded anymore (maybe release an upgrade from AOO to LO?). I also hope the OpenOffice trademark won't be allowed to lapse, as that would make it very difficult to combat malware-infected OpenOffice fakes.
Posted Sep 4, 2016 8:41 UTC (Sun)
by tialaramex (subscriber, #21167)
[Link] (6 responses)
So there's the seed of your problem right away. These "big committers" were actively hostile to the Apache project's choice of licensing. They actually wanted share-alike rules, and Apache is strictly opposed to that. Counting them as an asset was a grave error _by Apache_
Right there in the threads you linked is already the terrible sign that large numbers of people who actually _care_ about this stuff are giving a -1 against Apache's plan with long reasoned thinking and almost all citing the fact that TDF is the right home for this software. I don't know, maybe it's normal at Apache for all potential incubations to see such levels of disagreement ? Seems very unhealthy to me.
The other thing one or two people bring up that you've largely shaded out is that most incubations are for projects whose community comes to Apache and says we want to join. That community might be small and literally every member wants incubation, or it might be large and have come to this decision by some vote or other method. But what was rare (unprecedented ?) was for a new project to be "donated" as OpenOffice.org was, as a baby left on a doorstep. "Good luck, bye".
Posted Sep 4, 2016 12:26 UTC (Sun)
by jimjag (guest, #84477)
[Link] (5 responses)
>These "big committers" were actively hostile to the Apache project's choice of licensing. They actually wanted share-alike rules, and Apache is strictly opposed to that.
Nope, these big committers were people actively open and pro permissive licensing. As I said, they felt that their contributions were being abused by LO for the sole and singular purpose of defeating AOO.
>But what was rare (unprecedented ?) was for a new project to be "donated" as OpenOffice.org was
Not at all.
>"Good luck, bye"
That is one of the goals of incubation and creating a new project. To help it build a community where one didn't exist before.
>that large numbers of people who actually _care_ about this stuff are giving a -1 against Apache's plan
You ignore all those posts of people on both sides who saw this as an opportunity as well.
In any case, it certainly shows that all those people who say that Apache was clueless about the job it faced are completely wrong. There was resistance and disagreement for sure; there was also opportunity and potential as well. In the end the optimistic side won out. And as I mentioned before, LO benefited greatly from AOO, not least of which was its ability to rebase on a ALv2 AOO codebase and to inherit the IP provenance (and whatever IP related to patents, etc) that resulted from Oracle's donation.
Posted Sep 4, 2016 17:46 UTC (Sun)
by niner (subscriber, #26151)
[Link] (1 responses)
Posted Sep 6, 2016 12:09 UTC (Tue)
by roblucid (guest, #48964)
[Link]
Posted Sep 4, 2016 19:33 UTC (Sun)
by cortana (subscriber, #24596)
[Link]
> Nope, these big committers were people actively open and pro permissive licensing. As I said, they felt that their contributions were being abused by LO for the sole and singular purpose of defeating AOO.
Projection much?
Posted Sep 4, 2016 19:58 UTC (Sun)
by paulj (subscriber, #341)
[Link]
Posted Sep 5, 2016 9:03 UTC (Mon)
by moltonel (guest, #45207)
[Link]
So they wanted permissive licensing (which means embracing the possibility of giving without receiving) but didn't want LO to avail of that. I can resolve this conundrum in only two ways: 1) they had a naive idea of what permissive vs share-alike means 2) they specifically didn't want LO to be counted as part of the "wider OO community" despite the fact that it already was the largest subgroup of that community. Neither option shines a very good light on these big AOO contributors.
LO didn't "abuse" the AOO license, they used it as designed. If AOO's intent was for people to share-alike, then they should have chosen a share-alike license. The AOO community could have encouraged LO devs to triple-license their patches (which would have enabled AOO to cherry-pick the same way that LO did), but it seems that they just grumbled non-constructively instead. Conversely, the LO community attempted to mend the community by creating TDF and inviting people in, but that (unsurprisingly) did not work out.
Lastly, you complain about some LO devs wanting to defeat AOO. Ignoring the tit-for-tat reactions, it's completely normal that most of the LO community wanted nothing to do with AOO : LO contributors are pragmatic people who wanted to get the job done, and a huge part of that was getting rid of the Sun/Oracle bureaucracy, which Apache barely improved upon. The frustratingly long time for AOO to get set up and do their first "no-op" release was another sign that Apache was not a good home for the OO IP.
Really, AOO's lack of success was a surprise to nobody outside the AOO community. Pretty much everybody outside of AOO has been waiting (patiently or aggressively) for AOO to die off since its inception (no hindsight needed). Five years on, we're still hoping.
Posted Sep 5, 2016 15:09 UTC (Mon)
by nim-nim (subscriber, #34454)
[Link] (1 responses)
That's pretty disingenuous, TDF creation (should I say incubation?) was a long time in the making. It didn't sprang suddenly fully formed, pretty much all its core constituents were known beforehand.
It would have been trivial to just poll anyone with an OO.o account, GOO.o commit or OO.o distro maintainership rôle to learn how the land lay. (And Oracle/IBM knew perfectly well that sacking existing SUN OO.o devs was hardly a good start for a new project).
Indeed the OpenJDF history shows SUN and Oracle were perfectly able to feel the water and embrace existing developer communities when they wanted to.
The truth is that TDF was pondered many time during SUN last OO.o years, it crystallized when SUN was sold and SUN/Oracle and IBM were perfectly aware of its future existence.
And then it would have been utterly trivial to wait for the end of the TDF paperwork or even speed it up a little with some corporate help.
No, the real reason is
> Oracle [IBM] wanted to have the OO codebase under a permissive license.
And there was no way *that* was going to happen outside a corp-only project. SUN had just about convinced every third party they needed some form of copyleft to protect themselves from corp diktats, and Oracle had a worse reputation than SUN. Again, any due diligence in polling existing contributors would have shown that.
Creating a competing project to TDF at this stage was a trainwreck in the making. The Apache could only ignore it by turning a blind eye to inconvenient facts. It proceeded nevertheless, either for foolish idealistic reasons, or because some Apache members wanted an OpenJDK/Apache Harmony revenge.
Problem is, in both cases the IBM sugar daddy defaulted.
Posted Sep 5, 2016 21:53 UTC (Mon)
by foom (subscriber, #14868)
[Link]
Posted Sep 5, 2016 18:24 UTC (Mon)
by Wol (subscriber, #4433)
[Link] (3 responses)
> Finally, some big committers to AOO got frustrated that they were contributing useful stuff to AOO which then got pretty immediately consumed by LO, when, at the same time, there was resistance (ranging from minor to extensive) from stuff going the other way. Feeling that TDF wasn't "playing fair" these committers stopped, leaving a vacuum in AOO.
And oh my, how this does show the importance of perception over reality!
Not that I understand the figures someone else quoted - is it just 7 AOO commits found their way into LO out of 150, or was it 2000, commits, or are those figures just for 2016 (in which case I must say I think they are very misleading ...), but it comes over pretty clearly that far fewer commits went from AOO to LO than people think.
Are the AOO devs looking at the "LO reviewed these AOO patches" figures and assuming they were all cherry-picked?
And at the end of the day, there is (on the LibreOffice side at least) ABSOLUTELY NOTHING stopping the AOO devs from cherrypicking LO code into AOO. The EXISTING LO licence permits it. The only thing stopping AOO cherrypicking LO code is idealism on the part of AOO.
Cheers,
Posted Sep 5, 2016 20:36 UTC (Mon)
by nix (subscriber, #2304)
[Link] (2 responses)
It is true, perhaps, that the final calculation should have been 10,000-odd versus 150 commits, rather than versus the 7 that were also relevant to LO -- i.e. LibreOffice is *only* perhaps a hundred times more active than AOO at this point. Obviously that makes AOO a worthy upstream after all!
Posted Sep 5, 2016 20:48 UTC (Mon)
by Wol (subscriber, #4433)
[Link] (1 responses)
It's always been considered obvious that as time goes by, the amount of cherry-picking from AOO into LO will go down as the code bases diverge ...
Cheers,
Posted Sep 6, 2016 10:22 UTC (Tue)
by Felix (guest, #36445)
[Link]
Well OpenHub's code crawlers seem to experience some problems with analyzing the latest AOO code but anyway: I guess their graphs give a basic overview https://www.openhub.net/p/openoffice/commits/summary vs. https://www.openhub.net/p/libreoffice/commits/summary (restrict LibreOffice graphs to last 5 years to exclude the big spikes earlier).
As one can see there is not a single month in the latest 5 years where LibreOffice had less than 1000 commits (sometimes even 2-3k commits per month). In (roughly) the same time frame AOO saw only 4 *months* where it had more than 500 commits and basically since the fall of 2014 the commit frequency went downhill.
Posted Sep 6, 2016 12:07 UTC (Tue)
by roblucid (guest, #48964)
[Link]
You're kidding yourself if you think permissive licensing is about code-sharing, it suits companies who want to add their own secret "sauce" and create fragmented closed features.
Posted Sep 3, 2016 19:27 UTC (Sat)
by xtifr (guest, #143)
[Link]
Of course, as with any copylefted work, the AOO devs were welcome to approach individual developers and ask for permission to use their code in AOO. No matter how LO was licensed as a whole, the individual authors could relicense their code if they wanted. And I'm sure some would have said yes, especially if they didn't have to do the (by this time non-trivial) porting work themselves.
But of course, AOO, for a variety of reasons, lacked the manpower for any such effort. And had at least one very prominent member stirring up bad blood between the projects, making it less likely that individual developers would be interested in helping "his" project.
Really, AOO might have ended up in a much different place if they'd simply muzzled that guy early on!
Posted Sep 4, 2016 19:53 UTC (Sun)
by paulj (subscriber, #341)
[Link] (24 responses)
Further, clearly many people (given your "uni-directional" comment) *prefer* to *NOT* have to rely on altruism. Clearly, many people who believe in code being "given back" think the _most sensible_ way to achieve that is just to require it in the licence.
It is mystifying how supporters of permissive licences can simultaneously:
- Believe code must be available under a permissive licence, so (e.g., I guess) it can be used in proprietary code
- Get upset when other free software projects incorporate said code into copyleft projects.
Really, if you want people to follow certain rules or behave in certain ways with your code , just state those rules in the licence. And (unless people are exploiting some unintended loophole), don't get upset when people follow your licence - that's just illogical when _you set those rules_.
"They should have followed my beliefs, not my explicit licence!" is the weirdest kind of passive-aggressiveness in software.
Posted Sep 4, 2016 21:47 UTC (Sun)
by roc (subscriber, #30627)
[Link] (14 responses)
Posted Sep 4, 2016 22:43 UTC (Sun)
by DOT (subscriber, #58786)
[Link] (5 responses)
There is one role a proprietary project cannot fulfill: being a common base that everybody builds on, a place of open collaboration. Therefore, a proprietary product is never in direct competition with a permissively licensed project. But a (weak) copyleft project forked from the permissive project can take over this role. Moreover, the copyleft project then has an "unfair" advantage, in that it can take code from the permissive codebase while no code flows back[1]. The copyleft project then quickly looks like a more attractive place for collaboration.
So why go permissive, then, in the first place? Well, in theory there are some advantages over copyleft for businesses that want to create proprietary products. So if you have a strong community of businesses adding value on top of your project, a permissive license is a good fit. But these advantages mostly vanish in the case of LO vs AOO, because the MPL is just weak enough for most businesses. It also didn't help that there wasn't a strong proprietary community to fuel AOO development. If Oracle and IBM had thrived with their AOO derivatives, the outcome would have been very different.
[1] It is important to remember that there is nothing legally stopping Apache from taking LibreOffice's code and incorporating it with AOO. MPL and ASL are completely compatible. The problem is simply that Apache is not willing to take code with that license.
Posted Sep 5, 2016 6:24 UTC (Mon)
by paulj (subscriber, #341)
[Link] (4 responses)
Pretty simple really. Either they agree with their own licence and they should be happy when it's used in copyleft projects as much as proprietary, closed ones; OR they should use a licence that actually reflects what they want.
Posted Sep 5, 2016 14:31 UTC (Mon)
by Wol (subscriber, #4433)
[Link] (3 responses)
What you're describing is the default state of copyright - ALL the code in LO that is taken from AOO is still Apache - afaik the Apache licence doesn't allow you to relicence, it just allows you to commingle and distribute.
What I think you mean is a licence that is a "universal receiver" like the GPL - "if you mix Apache-AB with something else, the entire work must be distributable as Apache-AB". (with apologies to blood groups :-)
The problem with "universal receiver" licences is that they cannot be mutually compatible without horrible contortions.
Cheers,
Posted Sep 5, 2016 14:44 UTC (Mon)
by paulj (subscriber, #341)
[Link] (1 responses)
I.e., if they had wanted other open-source projects' modifications to be under their preferred licence, they should have used a licence that achieves that - assuming that "source code distribution" is a workable bright-line between the "open source project" they want to require code give backs from, and "proprietary, closed use" that they want to permit.
To paraphrase Rob Weird, title and licence is all that matters and don't complain if that doesn't suit you.
Posted Sep 5, 2016 18:30 UTC (Mon)
by Wol (subscriber, #4433)
[Link]
Cheers,
Posted Sep 5, 2016 14:47 UTC (Mon)
by paulj (subscriber, #341)
[Link]
Posted Sep 5, 2016 10:08 UTC (Mon)
by pabs (subscriber, #43278)
[Link] (6 responses)
I was told that OpenBSD do email campaigns to ask companies to give back source, but I can't find any reference to that.
Posted Sep 5, 2016 11:27 UTC (Mon)
by nhippi (subscriber, #34640)
[Link] (5 responses)
Posted Sep 5, 2016 20:59 UTC (Mon)
by Wol (subscriber, #4433)
[Link] (4 responses)
The copyright owner gave downstream the right to choose which licence THEY wanted to USE. He did not give them the right to change the licence he granted to their downstream.
Think of it as two separate parts - there is a big variety of licences out there - BSD, MIT, (L)GPL 2 or 2.1 or 3 etc. Note I very carefully did not say there was a licence called GPL2+, or any variant of plus. Because that is not a licence, that is a grant. It tells you which licence(s) you can use.
So if I grant you the right to use either BSD or GPL2+, that does not give you the right to take those choices away from your downstream. If you make a substantial edit to my code, and licence yours differently, that may change the terms on which the combined file may be distributed, but it does not change the terms I applied to my code. And deleting (or altering the meaning of) my notice is, as I say, imho a violation in itself.
Cheers,
Posted Sep 5, 2016 21:27 UTC (Mon)
by paulj (subscriber, #341)
[Link] (3 responses)
That may be a distinction you are unique in drawing, AFAICT. (?)
Posted Sep 6, 2016 15:24 UTC (Tue)
by Wol (subscriber, #4433)
[Link] (2 responses)
If you ever were unlucky enough to end up in court over this, you couldn't give the Judge a copy of the "GPL2+" licence. You would have to give him a copy of either/both GPL2 or/and GPL3, and the text that gave you the right to choose between them. And then the Judge would say, "well then, which choice did you make?".
Cheers,
Posted Sep 6, 2016 15:50 UTC (Tue)
by paulj (subscriber, #341)
[Link] (1 responses)
In the normal free software way, those terms are stated in each file, and refer to further terms described in other documents. Those terms can indeed give the licensee choices. E.g. to choose to use the GPLv2, or to use some later published licence by the FSF. Even the GPLv2 text within itself contains "either X or Y or Z" terms.
But, I don't see where you get that one of these descriptions is a "licence" and some other is a "grant"? I think you're mixing up words. A licence is granted by the rights-holder(s), and is done by describing the terms of the licence granted in some sufficient way - and those terms may have conditions, choices and refer to further documents with further terms.
Posted Sep 6, 2016 17:09 UTC (Tue)
by paulj (subscriber, #341)
[Link]
Posted Sep 6, 2016 15:31 UTC (Tue)
by MarcB (guest, #101804)
[Link]
A copy-lefted fork ruins this idea, because it turns the simple, straightforward act of adding features into a trade-off.
Posted Sep 5, 2016 15:13 UTC (Mon)
by orcmid (guest, #74478)
[Link] (7 responses)
There are occasions when cooperation is better than that, as in security matters, although sometimes the security-issue sharing can be clumsy. So far, missteps have been cleaned up.
There have also been cases where folks have contributed material from LibreOffice, but it is found that they do not have the right to do so and AOO reverts those contributions. The project also discourages anyone cherry-picking LibreOffice code and has intervened when that happened.
Finally, the ASF does not have a copyright transfer from Oracle. Oracle provided a license grant that allowed the ASF to release under its license. Oracle retains the copyright. Similarly, developers license their contributions to the ASF, they do not transfer copyright.
Posted Sep 5, 2016 16:11 UTC (Mon)
by spaetz (guest, #32870)
[Link] (6 responses)
Yes, but that is normally done under either the Apache (or a proprietary) license. What is different here is the the LO community prefers copyleft! And by triple-licensing they would undermine their preferences. So you really are angry that LO prefers to stick to copyleft principles. And that is not because LO devs are inherently enemies of AOO but because they havr a different worldview. Just like the Apache community sees value in a permissively licensed office alternative, they see value in a copyleft alternative.
As Jonathan Corbet said though, rehashing year-old arguments is not going to help here, so AOO should think how they get out of the hole they are in now.
Posted Sep 5, 2016 16:33 UTC (Mon)
by johannbg (guest, #65743)
[Link] (5 responses)
Nobody want's to see that Oracle plague infested spaghetti-codebase of an leviathan crawling back out of it's hole ( unless you are IBM ).
What people are wanting to see is for ASF to man up, shot it in the head then light it on fire and bury it's remains in the hole it already has dug itself and crawled into. ;)
Posted Sep 5, 2016 17:59 UTC (Mon)
by spaetz (guest, #32870)
[Link] (4 responses)
Posted Sep 5, 2016 18:57 UTC (Mon)
by johannbg (guest, #65743)
[Link] (3 responses)
Posted Sep 5, 2016 19:40 UTC (Mon)
by spaetz (guest, #32870)
[Link] (2 responses)
One last comment before I continue ignoring your future posts (you did go into my ignore file). I do object to your impolite and offending bile. Being followed by an emoticon does not make it any less offending and impolite!
I do not know how you come to think I am loyal to OpenOffice, my first commit into the LO repository was on 28-Sep-2010 for what its worth, when was your first constructive contribution besides offending people?
But whatever your stance on LO and AOO, treating people who invest voluntary time with respect is the very least that one can do. One can disagree, one can argue but sentences and analogies like yours are not something that one should tolerate in any civilized community.
Posted Sep 5, 2016 21:06 UTC (Mon)
by Wol (subscriber, #4433)
[Link]
Are you by any chance related to the famous (in psychiatric circles at least) Dr Albrecht Paetz from AltScherbitz?
Cheers,
Posted Sep 6, 2016 0:44 UTC (Tue)
by johannbg (guest, #65743)
[Link]
How I communicate is something I learned after series of events contributing my free time to a community through period ( of rather thankless ) 8 years it's not how I was when I started wet behind my ears but I quickly learned and adapted and build an immunity up to a certain point.
You can see this "dominant" communication method reaching all the way up to the highest level of the linux ecosystem, the kernel community with those being the individuals that set the example and tone for the rest of the linux ecosystem as an role model whether they like it or not.
How people respond,react and perceive communication is based on the environment in which they were raised which shapes their personality thus their "feelings" hence communication can never be "politically correct" no matter how hard it's tried but that does not prevent people from judging them (the perceived role models) as either good or bad ( something which does not exist ) based on that persons own perception which was shaped by that individuals environment.
If you remove the form of communicating from series of word written in text with video/audio instead you will see a completely different behaviour pattern in people and another one if people are met directly in persons.
That said contributing your free time to a community of any kind not just opensource or software in general is thankless work and will continue to be so until children are taught to put value on their own free time hence will start respecting others free time as an result of that.
However in a world driven by greed that's an effort that precisely will be prevented from happening as is being done already for a lot of man made problems which are often associated with the ( incorrect ) term of "saving the planet" when in fact "saving the human race" is the correct one.
Do you perceive the world you are currently living in with the rest of us as being "civilized" or ever been civilized in it's history ?
Posted Sep 8, 2016 8:53 UTC (Thu)
by ceplm (subscriber, #41334)
[Link]
It was **L**GPL for $DEITY sake! Nothing in the world stops IBM (or whoever) to take a whole list of libraries in LO and use them in their ReallyPriceyCMS or whatever. They would just have to help maintain that part of LO they use, which they apparently were too cheap to do.
Posted Sep 5, 2016 14:48 UTC (Mon)
by orcmid (guest, #74478)
[Link] (2 responses)
So, although this was just a device to allow a different license on LibreOffice distributions, it in fact put LibreOffice downstream from Apache OpenOffice, and that appears in their various notices.
Posted Sep 5, 2016 18:02 UTC (Mon)
by tialaramex (subscriber, #21167)
[Link] (1 responses)
If the actual upstream of an important project was in trouble, the downstream would have to choose between fixing it and taking over. But as far as LibreOffice is concerned, since they aren't downstream of you, this isn't a problem for them at all.
Posted Sep 5, 2016 18:37 UTC (Mon)
by Wol (subscriber, #4433)
[Link]
So if you're talking family trees, LO is more like AOO's aunt, not child.
Cheers,
Posted Sep 2, 2016 23:31 UTC (Fri)
by simosx (guest, #24338)
[Link]
Why does openoffice.org mention only AOO and not LibreOffice as an equal fork?
Several AOO contributors come off as condescending against LibreOffice. They probably do not even realise it. Even in the discussion at dev@aoo there is some condescending attitude. Such a thing is unacceptable, and probably contributed to the bad situation for AOO.
Posted Sep 3, 2016 13:22 UTC (Sat)
by dtardon (subscriber, #53317)
[Link]
AOO has done nothing to attract developers: there is no tooling, no infrastructure... LibreOffice has gerrit, continuous integration, regular Coverity builds, regular import tests using a large set of documents, etc. AOO can't even do builds. That's not mentioning all the red tape one has to go through to even become a contributor...
You have never explained how do you envision such sharing without putting all the burden on the consumers of the "common core". We are not talking about changed branding here. Nor is AOO a library/module/whatever that would be only used as a black box. That means that the structure of the consumer projects would be the core repo + a set of patches above it, in whatever way (for illustration, the current diff between AOO and LibreOffice has 14662481 lines and 609 MB). Well, we already seen this: it was called go-oo. And I can well understand why the people who remember it don't want to go back. But again, you've never asked these people about their opinion...
Btw, could you name the projects that are actually interested in such a "common core"? So we don't talk about hypotheticals.
Posted Sep 2, 2016 14:43 UTC (Fri)
by spaetz (guest, #32870)
[Link]
May I commend you for your public suggestions on the AOO mailing list? I found them sensible and level-headed despite the fact that whatever decisions will be taken (or not), a part of the community will always be unhappy and loudly complaining.
I found one thing especially remarkable in your post: whatever happens, the release of OO.org code under the Apache license was an achievement which should not be forgotten. Libre Office's building on that relicensed code was enabled by AOO.
The continued existence of a stagnant AOO offering insecure downloads however is a harm to the reputation of all Open Source Office suits. (although one could argue that e.g. Abiword is just as big an offender here :-)) And I am glad that the Apache board has noticed and is attempting to remedy the situation. Good luck with whatever path will work for you.
In case of retirement consider donating the OpenOffice(tm) to continue to make use of that brand (although I am not sure that LibreOffice would even consider rebranding at this point).
Posted Sep 5, 2016 19:58 UTC (Mon)
by nix (subscriber, #2304)
[Link] (1 responses)
The Apache Way is a process which, while not *necessarily* so, can easily be turned into a bureaucratic hellscape if enough contributors want to do so, more or less regardless of the views of any other contributors -- and there are now multiple documented cases of projects moving out of Apache because of certain people who shall remain nameless who migrate from minor project to minor project so that they can get onto their PMC by making trivial contributions for a while, apparently because they bizarrely think being on the PMC is an "important role" -- and then proceed to use that position to try to make the project "more Apache" by trying to force the project to use every form of voting and bureaucracy they possibly can, without considering whether this is actually sensible for the project in question.
In my eyes, this is only distinct from large corporate processes because nobody's getting paid -- oh and also they don't have the argument that huge numbers of people depend on the thing to justify their obstructionism either. (A good rule of thumb is that someone on your PMC is attempting to introduce more e.g. pre-release voting than Apache HTTPD does, that this is probably unjustified! I *wish* this was a joke.)
Posted Sep 5, 2016 20:26 UTC (Mon)
by nix (subscriber, #2304)
[Link]
If it was felled some other way, we should figure out what that is, rather than endlessly rehashing the same tired old all-the-libreoffice-contributors-are-selfish-noncontributors lines yet again. As Jon has noted, that's the past. Nothing can save AOO now, but it would be nice to know why it died (or why it never took off). Perhaps it was simply that everyone hated Oracle, but, y'know, btrfs was funded by Oracle for years and didn't die. It seems more likely that something Oracle or Apache *did*, or didn't do, kept contributors away. (The most plausible is simply the creation of a new project with a deliberately incompatible license when a perfectly good project based on the same code and with an already-lively contributor base already existed and had for years.)
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Noel Grandin noted on Hacker News that he tried to contribute to AOO, but ... there aren't people at AOO to handle the incoming patches.
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Wol
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Wol
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
almost all commits have a git-annotation...
prefer: <sha> which indicate that the said patch has already been addressed previously in master (although not necessarily the same way, and sometimes years before, including before the AOO fork started)
or
ignore: <various reason>
the most common reason being that the patch is aoo-specific, like changing version number and other house keeping
or that the patch is obsolete (like patching the ancient and long replaced dmake build system)
it is a bit like going to New Orleans, peeing in the river and declaring yourself 'upstream' of the Mississippi
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Wol
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
o https://lists.apache.org/thread.html/fbd76edce2746a7263f6...
o https://lists.apache.org/thread.html/7ae4d9224b6a4e3fa249...
o https://lists.apache.org/thread.html/e3e63d8d1a6e4c08ab1c...
Contemplating the possible retirement of Apache OpenOffice
How AOO ended up here
How AOO ended up here
How AOO ended up here
How AOO ended up here
How AOO ended up here
How AOO ended up here
How AOO ended up here
> Nope, these big committers were people actively open and pro permissive licensing. As I said, they felt that their contributions were being abused by LO for the sole and singular purpose of defeating AOO.
Contemplating the possible retirement of Apache OpenOffice
> which is that TDF didn't legally exist at the time.
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Wol
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Wol
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Wol
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Wol
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Wol
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Wol
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
*plonk*
Contemplating the possible retirement of Apache OpenOffice
This was known outcome the day that Oracle touch it and later dropped the ball on the community when it already had left.
This endless rerun of OpenOffice inevitable demise and the melodrama that surrounds OOo and how it has fallen into despair, disrepair, and relative abandonment that as been filling the internet since twenty eleven is long passed it's due.
Just take it into the backyard and out of it's misery before StarOffice legacy gets disgraced even further than it already has.
Accept the reality for what it is and be proud knowing that the legacy that now carries onward and lives in LO once was StarOffice.
Contemplating the possible retirement of Apache OpenOffice
*End of communication*
Contemplating the possible retirement of Apache OpenOffice
Wol
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
upstream
upstream
Wol
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice
Contemplating the possible retirement of Apache OpenOffice