4 ms·

"One possible solution would be for the OpenOffice folks to copy the Linux kernel and become an "upstream project" that others use to create a range of office-s

by jimjag 10y ago

"One possible solution would be for the OpenOffice folks to copy the Linux kernel and become an "upstream project" that others use to create a range of office-suite distributions like LibreOffice." ... which were exactly the thoughts and reasons behind Apache accepting OO in the 1st place. For various reasons this obviously never happened. But it WAS the idea from the start.

cesarb 10y ago
Interestingly, that could have happened. Not only did LO make the effort of rebasing all their changes on top of a particular AOO release, they also were set up to incorporate all the new work from AOO into their branch. For instance, they quickly incorporated the sidebar code. Even today, one of the LO developers still goes over each AOO change, merging any worthwhile ones. What actually happened? For some reason, development on AOO stalled. What little work still happens on it, is almost always replicating work LO already did two, three, sometimes five years ago. See it for yourself at https://cgit.freedesktop.org/libreoffice/core/log/?h=aoo/trunk https://cgit.freedesktop.org/libreoffice/core/log/?h=aoo/tru... : pick a commit, and look for the "Notes" field. For instance, picking "Port main/fileaccess to gbuild" I see the note "prefer: 1f9bc2b2aa8168f9c164044058b117d2a17d83ad"; that git hash is for a LO commit from 2011 (https://cgit.freedesktop.org/libreoffice/core/commit/?h=aoo/trunk&id=1f9bc2b2aa8168f9c164044058b117d2a17d83ad https://cgit.freedesktop.org/libreoffice/core/commit/?h=aoo/...). In the meantime, LO did a lot of cleanup work. Finishing the build system move to gbuild, cleaning warnings across all the code base, removing plenty of obsolete or unused code, running several static checkers, and more. As I noted at https://lwn.net/Articles/699108/ https://lwn.net/Articles/699108/, the recent AOO CVE was probably found and fixed on LO two years ago, as a result of running a fuzzer together with valgrind. If you wanted to pick an "upstream project" to base an office suite on today, it makes more sense to pick LO instead of AOO.
- jimjag 10y ago
  I think you are proving my point... AOO was "plundered" by LO but LO never gave anything back. So any "co-opporation" was extremely one-sided and one-direction. It was hoped that LO, as good FOSS neighbors, would donate back, even though they didn't have to, just as we hope that our corporate neighbors also donate back. We trust in altruism rather than in forcing it. It just so happens in this case that our trust was somewhat misguided.
  - davidgerard 10y ago
    Elsewhere in this thread, grandinj of LO notes that he tried to donate back and there weren't enough people at Apache to do anything with the patches. https://news.ycombinator.com/item?id=12414681 https://news.ycombinator.com/item?id=12414681 So you did get your wish, but it's not the LO people at fault.
    - jimjag 10y ago
      I'd like to see some of those patches...

by jimjag 10y ago

cesarb 10y ago

Interestingly, that could have happened. Not only did LO make the effort of rebasing all their changes on top of a particular AOO release, they also were set up to incorporate all the new work from AOO into their branch. For instance, they quickly incorporated the sidebar code. Even today, one of the LO developers still goes over each AOO change, merging any worthwhile ones. What actually happened? For some reason, development on AOO stalled. What little work still happens on it, is almost always replicating work LO already did two, three, sometimes five years ago. See it for yourself at https://cgit.freedesktop.org/libreoffice/core/log/?h=aoo/trunk https://cgit.freedesktop.org/libreoffice/core/log/?h=aoo/tru... : pick a commit, and look for the "Notes" field. For instance, picking "Port main/fileaccess to gbuild" I see the note "prefer: 1f9bc2b2aa8168f9c164044058b117d2a17d83ad"; that git hash is for a LO commit from 2011 (https://cgit.freedesktop.org/libreoffice/core/commit/?h=aoo/trunk&id=1f9bc2b2aa8168f9c164044058b117d2a17d83ad https://cgit.freedesktop.org/libreoffice/core/commit/?h=aoo/...). In the meantime, LO did a lot of cleanup work. Finishing the build system move to gbuild, cleaning warnings across all the code base, removing plenty of obsolete or unused code, running several static checkers, and more. As I noted at https://lwn.net/Articles/699108/ https://lwn.net/Articles/699108/, the recent AOO CVE was probably found and fixed on LO two years ago, as a result of running a fuzzer together with valgrind. If you wanted to pick an "upstream project" to base an office suite on today, it makes more sense to pick LO instead of AOO.

jimjag 10y ago
I think you are proving my point... AOO was "plundered" by LO but LO never gave anything back. So any "co-opporation" was extremely one-sided and one-direction. It was hoped that LO, as good FOSS neighbors, would donate back, even though they didn't have to, just as we hope that our corporate neighbors also donate back. We trust in altruism rather than in forcing it. It just so happens in this case that our trust was somewhat misguided.
- davidgerard 10y ago
  Elsewhere in this thread, grandinj of LO notes that he tried to donate back and there weren't enough people at Apache to do anything with the patches. https://news.ycombinator.com/item?id=12414681 https://news.ycombinator.com/item?id=12414681 So you did get your wish, but it's not the LO people at fault.
  - jimjag 10y ago
    I'd like to see some of those patches...