LWN: Comments on "Contemplating the possible retirement of Apache OpenOffice"

Contemplating the possible retirement of Apache OpenOffice

pabs — Sat, 14 Nov 2020 02:14:35 +0000

You should remove AOO and install and use LO.

Contemplating the possible retirement of Apache OpenOffice

allenwrench1 — Sat, 14 Nov 2020 00:43:05 +0000

Is AOO retired or is LO the new OOorg, im confused on which one I should be using...

Contemplating the possible retirement of Apache OpenOffice

flussence — Sat, 08 Jul 2017 23:24:11 +0000

Hmm, that's a good point with SNI. I spent far too long myself doing trial-and-error to make sure random bots hitting my server via IP didn't get a free redirect to a real vhost.

Contemplating the possible retirement of Apache OpenOffice

tialaramex — Fri, 07 Jul 2017 16:28:16 +0000

There are a whole bunch of problems, but the most central is Usability: It's not enough that in theory it's possible to configure HTTPS correctly with your software, here's a study where they let people who'd definitely be able to get an HTTP server working try to set up HTTPS with Apache, and er, it didn't go so well

https://www.sba-research.org/wp-content/uploads/publicati...

As well as things that study digs into, like Ciphersuite configuration there are individual pain points that study doesn't touch on such as:

OCSP Stapling: To get from "That's a nice protocol, shame it's so impractical" to a working revocation system we need OCSP stapling. Apache notionally supports it. So you'd think we could quickly get to, say, 50% install base. But their implementation was either built by someone who understood the individual words in the specification but had no idea what the overall goal was, or it's been intentionally sabotaged to put people off. Most people who try to set this up will shoot themselves in the foot and end up just making their server needlessly unreliable. Experts, with a LOT of effort can make it sort-of more or less work. That's nowhere near good enough for a feature that other servers get right out of the box.

SNI default handling: It's all very well to implement SNI, but what in particular should happen if the client requests a name we've never heard of? By default Apache's response is to shrug and press on regardless, as it would with ordinary VHosts. This opens up a number of unpleasant surprises, which in practice we now have to put up with because Apache is so widespread, but it'd be wrong not to _point out_ that Apache played a role.

Contemplating the possible retirement of Apache OpenOffice

flussence — Wed, 05 Jul 2017 02:10:10 +0000

Now I'm curious. They have SNI, ALPN/H2, chacha20 works out of the box with libressl, and certs reload on sighup. That sounds like a full-featured HTTPS server, what am I missing?

Contemplating the possible retirement of Apache OpenOffice

tialaramex — Mon, 03 Jul 2017 19:51:48 +0000

Three more months later, AOO's (Windows) buildbots have been broken for several weeks.

The PMC writes, after observing that they have been "waiting since some weeks" for the person notionally managing the 4.1.4 release to er, do any managing, without any trace of irony, "Apache OpenOffice 4.2.0 is planned for this year - but without to name a specific time frame" as if by power of wishful thinking "this year" isn't far too definite a time frame for the project to commit to.

These meagre resources should have been directed to winding things up in an orderly fashion, preferably years ago, but certainly by the end of 2016. Instead the Apache Software Foundation is allowing AOO to rot from the inside and its key personnel are in a state of denial.

Apache's pretty 2017 "annual report" cites Apache OpenOffice repeatedly, it doesn't mention that the project is largely moribund and nobody should be using this software because it's essentially unmaintained, instead it just keeps talking about how many people are downloading and using it, at one point asserting a value of _millions of dollars_ to the Foundation giving away this software over FY2017.

I think the belief in some circles has been that since Apache is such a "light touch" organisation, the contamination won't spread so it doesn't matter. If one, or ten or a hundred projects are allowed to rot rather than being aggressively retired once they're not viable, well, that's somebody _else's_ project, not mine, I don't care. But actually it does matter because the "Apache Way" is not really the bullet point list that the board loves to cite, but instead the actual practices of the organisation. And those practices have been allowing projects to rot and fall apart, they're going to do the exact same thing to projects you _do_ care about.

With a different hat on I care about a very different but no less famous Apache project, httpd. Apache httpd does HTTPS very poorly. When most people's HTTPS implementations were pretty crappy this didn't feel like a big deal. It's just another optional component of the httpd that's not very good, just don't use it. Put a reverse proxy in front. Or don't use HTTPS. But "don't use HTTPS" is not really a practical option in 2017 (though you wouldn't think that if you talked to Apache's own infrastructure people), and if someone is going to buy separate components to handle the HTTPS termination, you start to wonder what the HTTP server is actually _for_ in this picture, it's not talking to clients, it's not doing most of the actual work... you would think that concern about HTTPS quality would intensively exercise the Apache PMC and be the focus of much new work. You would be wrong.

Contemplating the possible retirement of Apache OpenOffice

flussence — Sat, 15 Apr 2017 22:09:47 +0000

I spent a while skimming through their mailing list just now.

It's giving me a “religious cult” vibe - they strain to never speak of the elephant in the room while they cheerfully polish the holes in the drywall. And if someone puts them in a position where they're forced to acknowledge reality (e.g. one long thread asking why there's so little user involvement), it devolves into angry defensive rants about preserving the purity of their bloodline, how LO isn't "real", etc etc. There are a few personal attacks from one person who thinks LO is an illegal venture and Rob Weir is a saint (they've obviously never read any of his comments on here), but there's an awful long chain of +1 replies to them afterwards.

Can't possibly fathom why their new recruits per year remains epsilon, with such a warm and welcoming culture... /s

Contemplating the possible retirement of Apache OpenOffice

tialaramex — Sat, 08 Apr 2017 09:59:30 +0000

Seven months later, AOO has... shipped a micro release containing the fix for CVE-2016-1513, and almost nothing else.

The _optimism_ alas remains much as ever, leading to their PMC writing in October...

"Apache OpenOffice 4.1.4 is contemplated for release around year end for further maintenance and available security fixes"

and then when that had failed to materialise in January...

"Apache OpenOffice 4.1.4 is planned for release in 2017 Q1"

and presumably the next quarterly report will say now they hope it'll happen in Q2. For a point release, something LibreOffice can ship about once per month without strain.

Imagine if you were a developer who had the mistaken impression that work for AOO would be much appreciated. You create a new feature, submit it to the AOO project in, say, June 2016 and... nothing. Is that feature in 4.1.3? No, that's just fixing a security bug from a year ago, is it going to be in 4.1.4 ? No, it's only a bug-fix release, need to wait for 4.2.0 which may be years away if it ever happens. Meanwhile you see a friend's feature for LibreOffice gets into a beta, then a final release, somebody finds a bug, the bug gets fixed, that's released too...

Contemplating the possible retirement of Apache OpenOffice

Wol — Wed, 14 Sep 2016 18:22:33 +0000

No. I think you will find that upgrades to Windows actively sabotaged WordPerfect.

Given that WP was the dominant word processor on DOS, do you *really* think it likely that M$ would ship an update to Win3.1 that would reliably break it?

Or ship an office suite that would reliably kill a pre-installed WordPerfect stone dead?

That's not anecdotes. That's personal experience. I could predictably kill WordPerfect by installing an M$ update. And as I say - they were the dominant word processor at the time. I can't believe that was an accident ...

And if you read the court case, it is very clear that M$'s communications with WordPerfect Corp over Win95 were absolutely full of porkies. "WordPerfect for Win95" relied heavily on functionality that was in the beta copies of 95, but was pulled for the release version! Which is probably why MS Office was so full of back doors - maybe it was this code that was moved into Office rather than Windows, or maybe the Office team were "encouraged" not to use the official API precisely because M$ senior management were planning this "bait and switch" for a long time. Unfortunately, Novell waited for the Netscape trial to finish before pressing their own claims, and got timed out. They should have opened the case and asked the Judge to toll it until the DoJ case finished.

Cheers,
Wol

Contemplating the possible retirement of Apache OpenOffice

JanC_ — Sat, 10 Sep 2016 01:07:45 +0000

Actually, WordPerfect was sabotaged by a combination of their own not-so-competent management and IBM's politics (which tricked them into investing a lot in WP for OS/2, instead of in a version for Windows); Microsoft had relatively little to do with that, except that they profited massively from WordPerfect's judgement error.

Contemplating the possible retirement of Apache OpenOffice

simosx — Fri, 09 Sep 2016 15:40:12 +0000

> What with the huge amount of "press" about this, the AOO has simply been overloaded w/ emails from developers and other contributors offering their help, skills, talents and support!
>
> Thanks to all!

Here are the list archives of the "recruitement" Apache OpenOffice mailing list,
https://mail-archives.apache.org/mod_mbox/openoffice-recr...

There is not a single contributor yet.

Contemplating the possible retirement of Apache OpenOffice

branden — Fri, 09 Sep 2016 08:01:27 +0000

One could, if one had no sense of proportionality.

Contemplating the possible retirement of Apache OpenOffice

HelloWorld — Thu, 08 Sep 2016 20:43:52 +0000

> Sorry for not having the time to locate such specific examples in Bugzilla.
I'm not asking for an example in Bugzilla, your argument may well apply there. Or maybe not, but either way, it's beside the point, I'm asking for an example in LibreOffice. And notice that lack of a sidebar is *not* a feature. If you don't like it, turn it off in the View menu. It's that simple.

Contemplating the possible retirement of Apache OpenOffice

flussence — Thu, 08 Sep 2016 20:17:04 +0000

>The Apache HTTPD Server is was is described as "legacy", losing market share from new software that is better designed.
Credit where credit's due: Apache httpd is still very relevant, because it's the *only* widely available web server with decent HTTP2 support in 2016. Nginx rushed their feature out crippled and incomplete so they could loudly proclaim they were the first, and lighttpd has less of a pulse than AOO.

Contemplating the possible retirement of Apache OpenOffice

Del- — Thu, 08 Sep 2016 18:21:26 +0000

> Netcraft has better free stats.

You are of course free to believe that. After following numerous statistics over the years, I have come to another conclusion. However, I am in no mood to spend hours documenting it.

Contemplating the possible retirement of Apache OpenOffice

simosx — Thu, 08 Sep 2016 14:15:09 +0000

To get a good picture with W3techs, you need to pay for their full report.
The general single chart that they provide is not helpful enough.

Netcraft has better free stats.
See https://news.netcraft.com/archives/2016/08/24/august-2016...

The fourth graph is about "Market share of the top million busiest sites".
It filters out any domain parking websites and websites that are not very active.
Normally the busiest websites are the oldest websites, which would favour any legacy server software that are painful to upgrade and switch.
However, the stats show a steady decline in the Apache HTTPD server.

Contemplating the possible retirement of Apache OpenOffice

jimjag — Thu, 08 Sep 2016 13:31:41 +0000

>Actually, it depends on where you get the statistics from. If you go by W3Techs, Apache is still in the fifties:
>https://w3techs.com/technologies/overview/web_server/all
>I believe Netcraft has a bias in its data, so at least I don't trust them more.

Yeah, most other surveys show a much different picture than Netcraft

Walking ghost

Wol — Thu, 08 Sep 2016 12:18:43 +0000

Hmmm

So all that cleanup work has been well worth it!!! :-)

Cheers,
Wol

Contemplating the possible retirement of Apache OpenOffice

paulj — Thu, 08 Sep 2016 10:47:53 +0000

Analogies have power to help one people understand some new thing superficially, however they do not have _predictive_ or logical power. An analogy is not an isomorphism. You can _not_ *make* an argument from an analogy (including inviting others to draw conclusions from one) - that would a fallacy.

Others pointing out that disconnect between the analogy and the actuality is therefore quite valid.

Contemplating the possible retirement of Apache OpenOffice

SukSuk — Thu, 08 Sep 2016 10:02:58 +0000

Sorry for not having the time to locate such specific examples in Bugzilla.

The Sidebar IMHO is a bad feature that should have stayed in Lotus Symphony. It is a shame that LO also chose to implement it. There is no point in having two similar office suites. Each suite should aim to be unique instead of being yet another clone.

I agree, no chance for OOo and LO to break the glass ceiling as long as M$ is a platinum sponsor to ASF and Google is advisory board member in TDF.

Contemplating the possible retirement of Apache OpenOffice

ceplm — Thu, 08 Sep 2016 08:53:50 +0000

> Believe code must be available under a permissive licence, so (e.g., I guess) it can be used in proprietary code

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.

Contemplating the possible retirement of Apache OpenOffice

tao — Thu, 08 Sep 2016 07:29:53 +0000

Well you know... Netcraft confirms, BSD is dying...

Contemplating the possible retirement of Apache OpenOffice

Del- — Thu, 08 Sep 2016 07:07:29 +0000

> Yeah, I knew that Apache httpd has been dropping due to nginx. Last I had checked, we were still at about 53%, but yup, I see it is now low-40s. Thanks for the update.

Actually, it depends on where you get the statistics from. If you go by W3Techs, Apache is still in the fifties:
https://w3techs.com/technologies/overview/web_server/all
I believe Netcraft has a bias in its data, so at least I don't trust them more.

Thanks for providing the ASF. It is the best effort on this planet to provide a breeding ground for permissively licensed software. Personally I am thrilled to see how ALv2.0 prevails over alternative permissive licenses (even with the resistance of some well-known software houses). I believe you have all the right motives, and I wish you all the best. I hope we can put the AOO situation behind us soon.

Contemplating the possible retirement of Apache OpenOffice

johannbg — Thu, 08 Sep 2016 05:56:18 +0000

So everyone the refute your claims of "created" here and ask for statistics to back your own claims up are trolls.
I'm not surprised of the rift between Apache Foundation and the rest of the world is as it is if their world view is like that.

And fyi this is about AOO and ASF not taking responsibility and resolving the situation that has been going on for past 5 years since clearly the AOO community has not been able to take care of it itself.

Contemplating the possible retirement of Apache OpenOffice

gstein — Thu, 08 Sep 2016 05:27:23 +0000

I don't understand what you're trying to get at here ... nearly ALL of the projects at the ASF started elsewhere. I can think of only a few of our hundreds that began their life at the ASF (beyond httpd, of course).

My intention with the "Apache name dropping" is that (IMO) the Foundation is creating a space for some great work that helps the F/OSS world and the software industry at large. It is not evil overlord that some commentators in this thread have implied. This also goes to my point about avoiding conflation between the AOO community, and the work of the Foundation (eg. people may be assigning various blame to the Foundation where it should rightly be placed upon the AOO community).

Mind you, I'm not here to throw AOO under the bus :-P ... I think the Foundation did exactly what it was built to do: foster communities that were interested in coming under its umbrella. And so we accepted AOO as one of the many projects of the Foundation. We all had high hopes for it, and maybe it will solve its current problems, and figure out a forward path.

I'm not going to dig up statistics for you. You can do that just as easily, on your own time. But let's stick to Apache Hadoop. Nobody used the term "Big Data" because the code to enable those processes was not *available* to them. Then Yahoo donated the codebase to the ASF, and Apache Hadoop was created. THEN people had access and the Big Data revolution began. (agreed that I misspoke: Apache Hadoop didn't create the *concept*, merely the basis for the *term* Big Data)

So from some internal code at Yahoo to a multi-billion dollar Big Data industry? Yeah. I'd say that is more successful. From a no-community engineering team, to a multi-national pool of contributors ... again: a huge win in community and development.

Pah. I'm gonna stop now. You're trolling. This LWN article is about AOO. I don't need to defend the Foundation and its projects, to you. You can choose whether you like the ASF. I choose to believe it is beneficial for the planet and the software industry.

Contemplating the possible retirement of Apache OpenOffice

johannbg — Thu, 08 Sep 2016 04:32:13 +0000

It's interesting to see how you continuously do "Apache" name dropping on project that did not start in Apache and even gone so far as claiming it CREATED concepts ( Hadoop is based upon previous work done by Google so give credit where credit is due )

Do you have any statistics showing that after projects have joined the ASF umbrella that they have become more successful and community surrounding them have grown stronger in terms of end user base and development power?

And since you mentioned Groovy how healthy was it for the project ( and it's sub projects ) having to move from github, the largest development base on the planet into Apache's own Git repository?

And on general note of all the projects in ASF how many are bit rotting and how many are succeeding?

Contemplating the possible retirement of Apache OpenOffice

gstein — Thu, 08 Sep 2016 02:24:39 +0000

> (Although those who suspect hostility towards the *copyleft*/share-alike movement may have some reasons for suspicion.)

Haha! Yeah, though my impression is that most OSS folks just shrug their shoulders at Free Software people. The latter tend to be more fervent because they want to ensure developers retain Rights to how code is used. The OSS folks don't really care about those Rights, so they remain unconcerned. ... But yeah, suspicions are valid :-P

And clearly the ASF as an entity only uses ALv2 because it has a specific intent and desire, to use a permissive license. But the Foundation doesn't evangelize its own position. Others can choose what works best for them.

> If the ASF ignores this mess and just goes on with business-as-normal

Oh, that won't happen. The Board is already involved, and that is never a good sign. Around the ASF, we generally describe the Board as a blunt instrument, rather than a precision scalpel. It just doesn't have the information necessary to take measured, cautious steps. So we blow up PMCs, replace Chairs, or other similar nuclear interventions. PMCs do *not* want the Board involved ... it rarely turns out well for them (in the short-term; in the long-, the interventions all worked and the new PMCs are healthy or closed).

> definitely *not* off-topic

Well, I will continue to respectfully disagree. I believe the topic here is what AOO will choose to do, moving forward. Not what happened in the past, or the relationship between AOO and the ASF.

Contemplating the possible retirement of Apache OpenOffice

gstein — Thu, 08 Sep 2016 02:10:11 +0000

> Not any more.

Yeah, I knew that Apache httpd has been dropping due to nginx. Last I had checked, we were still at about 53%, but yup, I see it is now low-40s. Thanks for the update.

> There was no other similar type of software (office suite that deals with end-users instead of server software) in the ASF, and the ASF did not have such experience.

I don't believe that. We deal with communities, not types of projects. We never had a database ... until Apache Derby. We never had a language ... until Apache Groovy. We never had a Big Data project ... until Apache Hadoop CREATED the whole concept of Big Data.

The AOO community needs to figure out what it wants to do. The Foundation will give them that space. We'll intervene if and when we find the community is no longer working on the problem.

The Apache Way

gstein — Thu, 08 Sep 2016 02:03:10 +0000

> Assuming the parent post was written Greg Stein, Apache board member:

Yup. That's me.

> But AOO doesn't formally exist either.

As a community it certainly does. The ASF is the corporate entity used to hold the community's intellectual property (specifically, the trademarks; I don't think the copyrights were transferred). The community and the corporation are separate. I just saw people getting confused by that. By your remarks, you clearly are not one of those, and you recognize how the Board provides some oversight for the community.

But the Board does not insert itself into the operation of the community. Even when they shoot themselves in the foot (eg. Rob Weir). *You* may want us to be more involved, but that is not how the Board/Foundation has chosen to operate in its 17 years of life. We don't believe it is our role to interfere, but instead to provide the legal framework, the basic rules of operation, some infrastructure, and help with (eg.) press and trademarks.

In any case ... the debate isn't about how the Foundation operates. Or at least I don't care to participate in that debate. I was responding to people misunderstanding that the AOO community is somehow "The Apache Software Foundation".

Re: security ... Of course that is a problem. That is the primary reason that the Board *has* stepped in, and pushed on the PMC to fix itself or close shop.

Why did we accept AOO? I ask, why not? You say "cuz somebody was doing better". That isn't our decision to make. What *we* saw was a codebase, some trademarks, and a small community that believed they could be successful. So we gave them a home. You can disagree with the choice, but I don't think it was inherently wrong. Many communities come to the ASF, asking to be part of our umbrella of communities. Why should we judge them, and turn them away?

The Board does not police the communities. That just doesn't scale, and it isn't the role of the Foundation anyways. The PMCs and their communities should be self-policing, per our guidelines (the Apache Way is the shorthand name). We worked with the AOO PMC for a long long time to get them to "fix" Rob, and for a while there, he was better. But I'm not gonna worry about the past. He's long gone.

The question is for the AOO community to decide what to do. They may ask the Board to close up their project, but we aren't going to make that choice for them.

Contemplating the possible retirement of Apache OpenOffice

lenov — Wed, 07 Sep 2016 23:00:53 +0000

+ 1 million. I know several administrations in France who switched to open source software ONLY because of OpenOffice and stuck with it even when LibreOffice took over. After a while, the status of AOO was such that these administrations reverted to a 100% Microsoft ecosystem.

Contemplating the possible retirement of Apache OpenOffice

HelloWorld — Wed, 07 Sep 2016 22:54:38 +0000

> Bugzilla has a lot to do with this.
Simply asserting that doesn't make it true.

> There are a lot of open issues to be implemented in the future. Issue A may request a certain GUI, issue B a different GUI...
Yes, so what? Again, here's the list of criteria for some feature F that would justify a fork:
- it must be impossible to implement in LibreOffice, be it for technical or social or some other reason. Otherwise, it's clearly best to simply implement it there and be done with it, since you'll otherwise end up with dozens of office suites, each of which implements a different subset of the features any given user needs
- it must be implementable in Apache OpenOffice in theory as well as practice. Otherwise both alternatives won't have it, making neither product better in that respect
If you cannot name such a feature, you failed to apply your argument to the situation at hand. And in fact, the only feature meaningful feature that AOO ever brought to the table, the sidebar, was rapidly ported to LO, making AOO redundant. That alone proves your argument doesn't fly *at all*.

> Look at the current M$ office market share...
The numbers alone don't mean *anything at all* as long as they're not accompanied by a theory as to why they are the way they are. In fact, it's easy to argue the other way around: MS's market share clearly indicates that interoperability issues, thin spreading of developer resources etc. have left the open source office suites unable to compete, thus resulting in a monopoly of MS Office. Now I'm not saying that's actually the case, but you'll have to provide some sort of argument as to why this assessment of the situation is so much more likely to be wrong than your's if you want to be taken seriously.

Walking ghost

shmget — Wed, 07 Sep 2016 22:32:11 +0000

yeah well there has been a lot of work and progress in the past 6 years:
ci build for windows, takes about 27 to 29 minutes.
the ci system is handling build on the 3 platform + a clang + plugin build for all incomming patch in gerrit.. and the slowest build is when the slave that get assigned for mac is one of the mac mini, which can take up to 1h40.
linux build time is much more variable depending on ccache hits rate. it range from 10 to 40 minutes...

a release build for mac, on a mac mini, building with all 100+ localization and signing, take about 3 hours.

between the 2 overlapping branches supported and the rc, the release train averages to one release per week. (these are on top of ci-build)

on top of all that there is also boxes that build what be call 'bibibisect' these build every commit of master and put the resulting binary in a git repo so that qa can do bisection on binaries. These bibisect repo, which we make one per 'epoch', which is about every 6 months worth of commit.. end-up containing 10-12K binary versions of the product for a weight of 8-12GB

even the daily clang+ubsan build takes 'only' about 6-7 hours

Contemplating the possible retirement of Apache OpenOffice

xtifr — Wed, 07 Sep 2016 20:53:06 +0000

I agree that the roles of the AOO team and the ASF in this debacle should not be conflated.

And you're right, those who think the ASF was deliberately acting to harm the Free Software community are pretty clearly insane. (Although those who suspect hostility towards the *copyleft*/share-alike movement may have some reasons for suspicion.)

At the same time, the fact that the ASF has done (and is doing) lots of good things does not absolve them of responsibility for whatever mistakes they may have made and/or are making. So the recitation of ASFs good deeds is not entirely relevant here, except as evidence against any anti-free-software motives.

> The Foundation has very little input (by design) into how the AOO community runs its project.

This may be something the Foundation will want to reconsider in light of these events. Especially since, it should now be clear, people may not always clearly distinguish the behavior of the ASF from the behavior of organizations operating under its banner. If there is no official code of conduct for ASF projects, then perhaps it's time to create one. If there is one, perhaps these events will suggest ways the CoC should be expanded.

If the ASF ignores this mess and just goes on with business-as-normal, then I think they will start to deserve more of the blame. Especially if this end up happening again, with another project.

> And, hey, if you believe that the ASF has a problem with its model, then feel free (tho it is off-topic for this article).

The model of how they run their child projects is definitely *not* off-topic.

Walking ghost

mathstuf — Wed, 07 Sep 2016 20:23:59 +0000

You know, using Ninja makes things *way* better on Windows. Is LO using a generator or is it naked makefiles?

Walking ghost

Wol — Wed, 07 Sep 2016 20:11:08 +0000

> Still, I agree with the general point - if you can't *start* a build within 24 hours and be confident in its success, you have a problem.

:-)

As someone who follows these things on the LO dev list, I agree with the poster who said it takes ages ...

I don't know the details, but iirc the main Windows build machine is an 8-core Opteron monster (it may well have been upgraded by now), and it takes a couple of days to build following a "make clean".

Anyways, the general comments are that Windows takes a LOT longer than nix, and my Athlon X-III (with 16GB ram) takes overnight to do a clean build on gentoo.

Cheers,
Wol

Walking ghost

mathstuf — Wed, 07 Sep 2016 12:58:24 +0000

If AOO does take 47 hours to build, imagine gaining an entire week's worth of time in your build cycle. Insane to think certain people were "build system work is worthless" when so much time could be saved with such work.

Contemplating the possible retirement of Apache OpenOffice

Del- — Wed, 07 Sep 2016 12:54:26 +0000

> I don't see how Apache OpenOffice is somehow supposed to have prevented open source desktop from succeeding in the last five years.

Then you cannot have spent much time on the front lines. Microsoft Office (together with specialized software and device drivers) has been the main barriers of entry for the desktop. For free software companies, the first natural target was replacing the office suite. Then (if successful) you could start looking into the full desktop. Problem is, everybody thinks of openoffice because of brand recognition, so their (those who are using alternative desktops) experience and perception is from AOO. That makes it very much an uphill battle, or more accurately a lost battle. Redirecting all those kicking the tires of open source office suites to up-to-date libreoffice would be a very important step forward for everybody who wants to see open source prevail.

Whether an open source desktop will succeed or not is another discussion, but it seems Google already has done it with Chromebooks. Of course, they needed a competitive office suite first, namely Google docs. I would even claim that was the hardest part of it.

Walking ghost

nix — Wed, 07 Sep 2016 12:48:43 +0000

Yeah, definitely. On my distinctly aged previous-decade 2GHz machines (with fairly fast spinning-rust disks and enough RAM to cache everything) a build takes about three.

Walking ghost

mathstuf — Wed, 07 Sep 2016 12:30:36 +0000

That build time probably the time of the ARM architectures which are much slower. The x86_64 builds took less than 5 hours.

Contemplating the possible retirement of Apache OpenOffice

simosx — Wed, 07 Sep 2016 11:48:55 +0000

> I would like to point those naysayers to the Apache HTTPD Server, which still serves over 50% of the websites on this planet.

Not any more. The market share of the Apache HTTPD Server for developer websites is continuously dropping and at the moment is down to 30%.
Between July 2016 and August 2016, the drop was a big 5.7%.

In the top 1 million websites (the important ones), Apache still has a market share of 43%, however it has been steadily dropping since 2010 with no sign of recovery.
In contrast, nginx is what is being used for top websites. The use of nginx is increasing at a faster pace and currently is at 28%.

The Apache HTTPD Server is was is described as "legacy", losing market share from new software that is better designed.

Source: https://news.netcraft.com/archives/2016/08/24/august-2016...

> Please feel free to comment upon the AOO community (and LO), and their problems. And, hey, if you believe that the ASF has a problem with its model, then feel free (tho it is off-topic for this article). But please stop conflating the two things.

The Apache Software Foundation has been taken for a ride when they accepted "OpenOffice.org". There was no other similar type of software (office suite that deals with end-users instead of server software) in the ASF, and the ASF did not have such experience.
The processes of the ASF are not suitable for such software.
It looks like the decision where OOo ends up with, was "anything other than The Document Foundation").
What is feeding the remaining people in Apache OpenOffice, is their personal hatred for LibreOffice.
This is a mess specifically for the Apache Software Foundation to deal with.