Red Hat dropping support for LibreOffice

Which will probably damage Fedora ... I thought the whole idea of Fedora was so that developers could run "Pre-RHEL" and find bugs. No office suite? No *supported* *usable* desktop? No developers.

Cheers,
Wol

For internal collaboration most things are Google Docs, inbound external documents are PDF, internal technical docs are Sphinx / Markdown / PlantUML with source in git and HTML / PDF published by a build server and LO is mainly for opening historical word documents.

I'm guessing* Red Hat has enough information about its user base that it sees LibreOffice use waning enough that it's no longer necessary to package it as part of RHEL.

Probably 365 or Google Suite, yes, for the small sliver of users who still use RHEL as a workstation / desktop OS and have any reason to fiddle with office-y files regularly.

Red Hat moved away from LibreOffice years ago, and I (anecdotally) see very little use of LibreOffice in the wild. The LWN audience is going to be an outlier here, but the days of everybody needing a desktop office suite with Microsoft format compatibilty(ish) are coming to a close.

Google's suite is "good enough" for a lot of people and it has far better collaboration features than Microsoft Office and LibreOffice. (I recently did a brief stint with a company that standardized on MSFT stuff, and Word's collaboration features are shockingly clunky in comparison, even in the online version.)

Companies are also picking up tools like Notion, Airtable, SmartSheets and other online collab tools that tie in with Slack and lessen the use of traditional office suites. The company I work for now uses Google Docs for some functions (actual spreadsheets, some document collaboration) but has largely moved to Notion for a lot of things that would've been done in Word-type document or spreadsheet. (If you have used Calc, Excel, Sheets or whatever for things like project planning or whatnot where it's used as an ad-hoc mini-database or simple collection of tables... Notion and such are far better.)

I used to be a heavy user of LibreOffice, but it's been years since I needed it regularly. When Red Hat moved to Google Suite or whatever it's called, I had very little use for it professionally. Most of the time when I've collaborated with people for other projects it's also been in Google Docs or whatever. The days of shuttling around office-formatted files via email are well in my rear-view mirror.

My kids (20 and 17) use Google stuff pretty much exclusively for their classes, maybe some actual Excel for the 20-year-old who'll be a college junior next year. My wife probably touches actual Office files a few times a year when helping to review a chapter or article for a colleague, but that's about it. Desktop-based office suites are legacy applications. I'm not aware of any really strong FOSS contenders in the online space.

*Disclaimer: Former Red Hatter here. Left last May, but I have no insider knowledge of the specific stats the RHEL team may have access to or any knowledge of the conversations that led to this. But I feel fairly confident this decision wasn't taken capriciously or without considering the number of users impacted. If LibreOffice was still in sufficient demand by RHEL users, I doubt the RHEL team would be deprecating it.

That is exactly what they do. Well, some. My users / company are. And they are happier than they ever were with LibreOffice. They always hated using a fileserver for sharing, the crude group setup there, not finding documents again / no search, and then they saw multi-user concurrent editing, and never looked back.

It made me sad from an open source point of view. But I can't argue with any of that, feature wise.

I think it would be a good way how to do it with Flatpaks, actually, perhaps, the only way which could work.

Dropping LibreOffice is not a good idea.

Cheers,
Wol

They do, but it runs in the browser, and it is either G Suite or Office 365.

Also, the browser typically is Chrome or Edge, not Firefox.

> Somebody please elucidate the reasoning behind this, 'cause I sure don't get it.

I'd argue local LibreOffice is now completely irrelevant in the business world. The collaborative features offered by the cloud providers are so far ahead, that it isn't competitive any more (and those features are used intensely).
The only version that could potentially still be relevant would be the Collabora version, but this is also running in the browser.

My company made the switch from Exchange + MS Office to G Suite four years ago. Since then, running a Linux workstation has finally become absolutely painless. *Everything* works perfectly. But besides the foundations, nothing on this workstation is free software any more.

Because, well, how much engineering does RH throw at LO?

There's two ways to read that, and I read it that RH was pulling away from support for LO. That reduces engineering resource available to LO. That damages LO.

The resource to package LO on Fedora/RH probably isn't a lot. The resource to make sure LO works well for users on RH/Fedora is a lot higher. Losing THAT resource will hurt.

Cheers,
Wol

Because that's what's going to happen?

>Just run the Flatpak version.

Why is RHEL shipping any program at all? They could just ship the kernel and flatpak and be done with it.
Or they could stop shipping anything and just tell people to download Windows instead. That's even easier.

Flatpak, and encouraging the reliance on it, is moreover an abdication of responsibility for your platform. Every package you encourage to just use a flatpak for is turning that area into Somebody Else's Problem. Any issues that arise are up to someone else to deal with. Which sounds great, but this is putting it in the hands of people who don't necessarily give a wet snap about your platform. Moreover, it's troubling that RHEL is in such a financial state that they feel they can't support native packaging of a major software suite like LibreOffice.

Flatpak utilization is also having the very unfortunate effect of supporting the balkanization of Linux distributions. They don't have to actually work together to commonize their installation mechanism, because Flatpak will just solve that problem.

So it's not just Flatpak to the rescue. Its use has costs in every aspect of a distribution.

https://wiki.archlinux.org/title/HDR_monitor_support

Unless it is Microsoft Word. Collaborative editing outside the browser. People who still use Google Docs are most likely victims of inertia since Google Docs came first; but since that time Microsoft had at last obsoleted them.

If Red Hat customers were demanding more contributions then that’s what Red Hat would be doing. The idea that Red Hat owes transparency in its use of funds is painfully naive. (Outside the requirements IBM has for its public filings as a publicly traded company, that is)

This strikes me as a very... dishonest description of the situation.

For better or worse, distributions do a lot more than "repainting the bikesheds". Replacing distribution packages with upstream binaries is not a zero-impact change — yes, "flatpaks allow apps to be built once and used across distros", but the user experience quality suffers as a result.

The fedora /usr merge was completed over a decade ago: https://fedoraproject.org/wiki/Features/UsrMove
> Targeted release: Fedora 17
> Last updated: 2012-02-07
> Percentage of completion: 100%

The debian folks, however, are /still/ spending time and resources trying to figure out the best path forward.

I don't view the actions of RH here as "antics," I view them as a simple "we don't have the time" statement. Don't get me wrong, flatpak isn't exactly perfect, and I would say that reviewing some of the email thread, they probably should have consulted with the fedora folks first (especially if someone can step up and be an appropriate packager). But at the same time, they made a decision and moved forward. And, who knows, maybe the direction will change based upon the ongoings. It's a little early to assume that this decision is currently set in stone.

You may disagree with the direction, with the approach, with the end result, and that's all fine. But I just can't fathom the inefficiencies involved in how the debian project actually gets work done. I would 100% prefer a less than ideal decision here or there, if the alternative is taking in excess of a decade to figure out how to pull off something like the /usr merge.

LibreOffice Flatpak is directly maintained by upstream as seen in

https://flathub.org/apps/org.libreoffice.LibreOffice

For deployment, you can either run your own yum repo (it's really not that hard) or use ansible/puppet/whatever-is-hot-this-week to automate it.