Posted: Sun Feb 05, 2012 12:49 pm Post subject: Upgrade to LibreOffice 3.3.x?
This is not my own feature request but really a question for all of the NeoOffice donors: should we start upgrading the NeoOffice code to use the LibreOffice 3.3.x code or should we focus our development efforts elsewhere?
NeoOffice 3.2.1 is based on the OpenOffice 3.1.1 plus Novell's Go-oo modifications and since Novell's Go-oo engineering group is now LibreOffice's core engineering staff, upgrading to use a recent LibreOffice version is the logical upgrade path for NeoOffice.
My rough estimate is that it would take only about 2 months of work to upgrade the code and put out a NeoOffice 3.3 Beta release. However, there would likely be lots of bug fixing work to do throughout the remainder of 2012 since, based on past upgrades to newer OpenOffice.org and Go-oo code, we will likely find many new crashing or feature bugs.
So the question for NeoOffice donors is whether or not there are new features in LibreOffice 3.3.x worth the time and frustration that donors will feel working through any new bugs that are found?
I think that by upgrading to LibreOffice 3.3.x we will have less new bugs than if we upgraded to their newer 3.4.x version due to the fact that LibreOffice 3.3.x is based on the last stable OpenOffice.org version that Oracle released before they stopped funding the OpenOffice.org project. Nevertheless, if we upgrade to the LibreOffice 3.3.x code, we should expect to find new significant bugs throughout 2012.
Are any of the donors in favor or against upgrading the LibreOffice code? I would like to hear how at least a dozen donors feel about this. In theory, if many donors are in favor of this, I would start work on this immediately after I finish the new patch download floating window.
I'm for it, there were a couple of useful features. _________________ "What do you think of Western Civilization?"
"I think it would be a good idea!"
- Mohandas Karamchand Gandhi
Posted: Mon Feb 06, 2012 10:54 am Post subject: Re: If not around LibreOffice 3.3.x or 3.4.x source, where?
grahamperrin wrote:
Are those two pages ideal for a semi-technical person like me to get a feel of which one might be preferable as a starting point?
Yes. But be aware that many of the LibreOffice 3.3.x new features are already in NeoOffice 3.2.1. For example, SVG import, 1 million rows in spreadsheets, and bundled extensions like presenter view are already in NeoOffice 3.2.1 as they were in the Go-oo 3.1.1.1 modifications that we used.
Also, note that many of LibreOffice's text layout and rendering new features are applicable only to Linux. The NeoOffice code uses full native Mac OS X rendering using Apple's CoreText functions whereas all versions of LibreOffice still use the Apple's older and now obsolete ATSUI functions to layout and render text.
I didn't see anything very compelling in those lists of features, and I already know from the WordPerfect group I follow that 3.4.x has a serious bug that prevents the import of WordPerfect documents (it's been fixed in 3.5 release candidates).
The other things I saw from the lists Graham linked were that they made lots of changes to behavior and keybindings "to better fit with Linux expectations"; did they #ifdef LINUX them, or apply them universally to all platforms? Also, did they remove the built-in help, or just add additional help menu items that point to their wiki? Having help always with me (built-in) is one of the great features of the current software I would hate to lose.
On the other hand, after donations, one of the most common "knocks" against NeoOffice I see is that "NeoOffice is so far behind OOo/LO in versions" (whether or not the new versions contain features they want/need; it's just the version number that matters ) And it's not realistic to stay on OOo/Go-ooo 3.1.1 forever.
So I'm on the fence; I'm not necessarily opposed to upgrading, but I also don't see anything compelling to me to put me in the "rah rah upgrade!" camp
Smokey _________________ "[...] whether the duck drinks hot chocolate or coffee is irrelevant." -- ovvldc and sardisson in the NeoWiki
I didn't see anything very compelling in those lists of features, and I already know from the WordPerfect group I follow that 3.4.x has a serious bug that prevents the import of WordPerfect documents (it's been fixed in 3.5 release candidates).
The other things I saw from the lists Graham linked were that they made lots of changes to behavior and keybindings "to better fit with Linux expectations"; did they #ifdef LINUX them, or apply them universally to all platforms? Also, did they remove the built-in help, or just add additional help menu items that point to their wiki? Having help always with me (built-in) is one of the great features of the current software I would hate to lose.
These 2 cases only add to the reason that I only proposed upgrading to the LibreOffice 3.3.x code and the thought of upgrading to LibreOffice 3.4.x or 3.5.x makes me very, very nervous. My feeling is that since LibreOffice's engineers (almost all of which are employed by Novell and RedHat) are primarily focussed on making improvements for Linux users.
Not that that is a bad thing, but Go-oo's past customizations that pleased Linux users sometimes proved very irritating and/or unstable for NeoOffice users and, as a result, I have had to remove many Go-oo modifications to avoid "Linuxifying" NeoOffice. The Linux features removal list is probably still manageable for LibreOffice 3.3.x since that version is based on a stable OpenOffice.org release, but I suspect that the list has grown (and will continue to grow) in LibreOffice 3.4.x and higher.
So, if donors really want any specific features from LibreOffice 3.4.x, I would recommend that we upgrade to the LibreOffice 3.3.x code and then try to backport the specific LibreOffice 3.4.x features requested by our donors. We have successfully used that approach before to fix many OpenOffice.org 3.1.1 file import and export bugs that OpenOffice.org fixed in later releases.
sardisson wrote:
On the other hand, after donations, one of the most common "knocks" against NeoOffice I see is that "NeoOffice is so far behind OOo/LO in versions" (whether or not the new versions contain features they want/need; it's just the version number that matters ) And it's not realistic to stay on OOo/Go-ooo 3.1.1 forever.
I suspect that this reflects the fact that the needs of NeoOffice and LibreOffice's core user groups are generally different. My feeling is that NeoOffice is mostly used by individuals, small business, or small work groups within a large organization.
While the type of users use LibreOffice, both OpenOffice.org and LibreOffice likely have a lot of deployments in large organizations since they have binaries for Windows, Linux, and Mac. In general, my experience is that IT staff in large organizations like consistent versions across different platforms in order to be able to provide consistent support to all of the internal staff.
To be honest, products like NeoOffice and iWork that are customized for just one platform are not something that IT staff in large organizations want to deal with so I doubt that NeoOffice will ever be well received by that user base. Hence, over time I expect that LibreOffice and NeoOffice will diverge in very different development directions.
Joined: Jun 11, 2006 Posts: 481 Location: Great Britain
Posted: Mon Feb 06, 2012 2:44 pm Post subject:
Another dubious looking feature of 3.4: "Substantially re-work legacy StarOffice (pre 2000) binary file format filters, removing all export code: the filters are now smaller and read-only, this also removes clutter from the 'Save As' dialog"
I would support upgrading to 3.3 if the work needed to get onboard with LibreOffice now will save time and effort later. I don't see anything compelling in the new features that would make it important to me other than as a strategic move.
I'm using NeoOffice mostly as a "Swiss army knife" to view the plethora of MS Office documents I'm sent that I need to be able to view (and sometimes edit) in order to do my job. The important things to me are broad compatibility, UI performance in general, and stability.
My main gripe with OOo is Calc, which is inferior to Excel. As such, the 3.4 Calc features are of interest, the rest I do not care about much.
For 3.3, the Calc and Impress features are interesting and the SVG import.
The latter brings me to another question: If go-OOo went Linuxy after the Big Fork to LibreOffice, why not go with the Apache codebase and add the go-OOo patches from when they were still unified around 3.1.1.
I can't say if they make better code, but the tone seems more about cleaning up and sorting out messes than scoring Linux points. Their implementation of SVG sounds quite sane and helpful as it takes away the problem we always had with not being able to export vector graphs.
This bit of lateral thinking might be making things more complicated here, but if there is a divergence of interest, it may make less sense to stick to the go-OOo/LibreOffice codebase. There is also no Sun/Oracle management to balk at anymore.
Patrick, you are the programmer so feel free to tell me I writing doodoo here . _________________ "What do you think of Western Civilization?"
"I think it would be a good idea!"
- Mohandas Karamchand Gandhi
Another dubious looking feature of 3.4: "Substantially re-work legacy StarOffice (pre 2000) binary file format filters, removing all export code: the filters are now smaller and read-only, this also removes clutter from the 'Save As' dialog"
FWIW, I think those are the .sd* formats, so two generations back; I don't know that I've ever seen one of those in my life save for ones I created when testing the icon set.
So, practically speaking, I don't think anyone would get worked up about that (as opposed to, say, removing Office 97/98/2000/XP export, which would definitely be a dealbreaker lots of places). It's nice to be warned that they're doing that, in case you're the 1 person in the world that depends on StarOffice export; however, I do agree that it's a rather lame feature to tout--it's more of a programmer-chest-thump: "Look at my geek credentials! I rewrote code, made it smaller, and removed features!"
Smokey _________________ "[...] whether the duck drinks hot chocolate or coffee is irrelevant." -- ovvldc and sardisson in the NeoWiki
The latter brings me to another question: If go-OOo went Linuxy after the Big Fork to LibreOffice, why not go with the Apache codebase and add the go-OOo patches from when they were still unified around 3.1.1.
LibreOffice 3.3.x is essentially what you describe. Novell's Go-oo engineers plus an engineer or two at RedHat are the core LibreOffice engineers so when LibreOffice was created, all of the OpenOffice.org 3.3.x code plus the matching Go-oo modications are what was put into the LibreOffice 3.3.x source repository.
ovvldc wrote:
I can't say if they make better code, but the tone seems more about cleaning up and sorting out messes than scoring Linux points. Their implementation of SVG sounds quite sane and helpful as it takes away the problem we always had with not being able to export vector graphs.
Ed and I are both extremely nervous about OpenOffice.org 3.4.x because all of Oracle's engineers stopped work when OpenOffice.org 3.4 was still a beta release.
Oracle staffed a fairly large professional software test team and still past OpenOffice.org beta and X.X.0 releases have been very buggy and it wasn't until the OpenOffice.org X.X.1 release that many of the new crashing bugs got fixed.
Since nearly all of Oracle's OpenOffice.org developers are gone and neither LibreOffice nor NeoOffice has the test staff that Oracle provided, I think that for the next year or two we are in uncharted territory so my gut feeling is that we should be cautious and pull in OpenOffice.org 3.4.x components that donors request by backporting one or two at a time and pushing out those backported 3.4.x features in NeoOffice patches so that, if any of those features cause a significant crashing bugs, we can focus on fixing those bugs before pushing the next set.
....It's nice to be warned that they're doing that, in case you're the 1 person in the world that depends on StarOffice export; however, I do agree that it's a rather lame feature to tout--it's more of a programmer-chest-thump: "Look at my geek credentials! I rewrote code, made it smaller, and removed features!"
I do not understand why Novell's engineers have been trying to get rid of those file formats. Granted, it makes the installer bigger, but once the installer is 150 MB or so, it doesn't make sense to me to try and remove 5 - 10 MB from to installer in exchange for making it impossible for those who have document repositories with old StarOffice documents to open their files. And as we discovered when that above thread, there are still organizations that have lots of old .sd* documents that they occasionally access.
Joined: Jun 11, 2006 Posts: 481 Location: Great Britain
Posted: Mon Feb 06, 2012 11:38 pm Post subject:
pluby wrote:
All of the Go-oo modifications as far back as OpenOffice.org 3.0 disables support for .sd* files by default.[...]
I do not understand why Novell's engineers have been trying to get rid of those file formats.
I wasn't sure if "pre-2000" was an allusion to MS 95/97. Thanks both for the clarification.
Although loss of .sd support isn't an issue for me, I agree it seems odd to remove features for so little gain. Several times over the years it has been a godsend to me to discover an application that was able to open/save some obscure long-abandoned file format, be it video, graphics, audio etc.
Perhaps Novell's engineers have gone all Jobsian and decided to start "skating to where the puck is going to be." Could it be we'll say goodbye to the floppy disk save icon?
Let us not go overboard on this one. They removed the export option. So that means you can stil import, you just have to save your edits in a more modern file format.
And saving 5-10 MB is not a bad thing. I'm sure Patrick can tell you those MBs add up.
But if it is too risky, don't do it - Apache or LibreOffice. _________________ "What do you think of Western Civilization?"
"I think it would be a good idea!"
- Mohandas Karamchand Gandhi
OK. We definitely got off-topic a bit but from the discussion I think we can conclude that LibreOffice's engineers are trying to go into very different direction than OpenOffice.org or NeoOffice has so I'll summarize the conclusions I have so far from the discussion:
- Upgrading to a recent OpenOffice.org or LibreOffice version will add some new features to NeoOffice that users will likely want.
- Upgrading to LibreOffice 3.3.x is relative safe but the number of new features is not very large. That LibreOffice version is based on the last stable OpenOffice.org release and it appears to only include the Go-oo modifications that are not experimental.
- Upgrading to LibreOffice 3.4.x brings a lot of changes but is relatively risky. The track record of the Go-oo project (the predecessor of LibreOffice) for maintaining stability in the code has not been very good and we have been stuck with fixing some very nasty crashing and data loss bugs of theirs. That, coupled with the fact that LibreOffice 3.4.x is based on an unstable, abandoned release of OpenOffice.org makes upgrading to this release highly risky.
Based on the above assumptions of risk, here is what I propose we proceed with development:
1. Upgrade to the latest LibreOffice 3.3.x code and get a Beta release out in April 2012 and a final release by May or June 2012. When the final release is out, NeoOffice 3.2.1 would become the free download release and NeoOffice 3.1.2 would drop off the download site.
2. Immediately after the final release is out, we start backporting popular OpenOffice.org 3.4.x and LibreOffice 3.4.x features into NeoOffice using the NeoOffice. We can use the New Feature Requests forum to have donors identify and vote on which features to backport first and, based on our past experience we should be able to backport many features and put out a patch with the backported code within a month or so after the feature is proposed.
Although the Mac OS X Lion release and working through a long list of deferred IT work on our servers consumed most of our time in 2011, most of 2012 should be available for us to do the above plan.
Does anyone think this is a bad plan or that I am missing any key information?
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum You cannot attach files in this forum You cannot download files in this forum
) And it's not realistic to stay on OOo/Go-ooo 3.1.1 forever.
.