You are fooling yourself if you think libreoffice has feature parity with ms office, and gimp with Photoshop. Gimp doesn't even have feature parity with krita.
I have never, in 20 years in the business world - and that includes long years working closely with people in controlling - see anyone do something in Excel that could not be achieved in LibreOffice with similar effort.
Yes, there is no 100% feature parity. But there is feature parity in the things that people actually use.
Yes, I think MS Office / Libre Office is a bad example. Adobe Cloud (Photoshop) to GIMP and Maya (or equiv) to Blender are far better points and counter points.
The Adobe Creative Cloud ecosystem has very strong lock in todo with workflow and collaboration (amoung other things). It's not that you couldn't do the vast majority of those things with another tool (open or not, e.g. Affinity), it's that the company you're going to work for is going to look at your Resume for experience with Adobe, and issue you a laptop with adobe on it, expect you to produce AI and PSD files (and now of course Figma for certain subsets).
GIMP is clunky and is not pushing the boundaries. Some of the newer smart tools in Photoshop are impressive in their ability to select and delete objects. Even Affinity which offers very polished products is going to struggle.
Blender on the other hand is a prime example of where an open tool can make strides to break the stranglehold. It has a very focused team of collaborators and a clear roadmap that is taking it from a tool that was quite hard to learn even if it was powerful into something that could become the industry standard.
The UX of Blender compared to what it was 5-10 years ago is so much better. I'm not sure the same could be said of GIMP.
That power comes both from strong direction and a growing and active community.
> - see anyone do something in Excel that could not be achieved in LibreOffice with similar effort.
> Yes, there is no 100% feature parity. But there is feature parity in the things that people actually use.
It's not close to being wholly about the capability of the product itself.
Excel has the incumbency advantage - more people are familiar with it and more people already have it. Switching to LibreOffice therefore often means switching to a tool you don't know as well and then living with interop issues with your co-workers that are still on Excel. For LibreOffice (or any other competitor) to do something with "similar effort", it has to overcome those headwinds.
If you look at the history of spreadsheets, there is a long history of "better" products coming along, but only a few that were good enough to be able to displace the incumbent.
* VisiCalc - The First
* Lotus 1-2-3 - Runs on the PC, has graphics, and a bunch more capacity than VisiCalc
* Excel - GUI support
* Google Sheets (maybe) - Online real-time collaboration, etc.
There were many others along the way, and many of them could make the same claim you're making here about LibreOffice. (Multiplan, Improv, WingZ, Quattro Pro, Framework, etc.....)
Parity is ok, but the goal really has to be 'so much better that it justifies all of the switching costs'....
To not deal with those headwinds because it is difficult to pick-up a new tool—especially if you haven’t in a long time or aren't interested in the tool/technology—we should priortize kids learning FOSS tools. Especially if it's not often used, I think most folks we be happy with both meanings of ‘free’ as something to pick up every once in a while. Most of the differences matter only once you're a power user. I use LibreOffice for the reason because I open up a document once a month, not my daily work.
And Krita doesn't hold a candle to darktable for working with photography on RAW images. So? The point is that the basic features a students need to know will be covered by free software or there will be an alternative that does. Having feature parity isn't important if copying another software isn't a goal--versus just making something good.
How many advanced GitHub features not present on other platforms do you think students are using? If the answer is none, then there's no reason to choose a proprietary service they can't hack on and are required create accounts with a for-profit entity.
I notice from your blog that you have links to a number of commercial services including Keybase. Keybase is another really great example of the disconnect between the underlying open source (GPG, etc) and something that is easily approachable and just lets people get on with the things they want to do.
We shouldn't use Keybase, but the nearest alternative open-source, Keyoxide, is not functional unless you're an engineer. Keybase on the other hand is perfectly usable without being an expert in the underlying cryptography.
Similarly Github has a huge number of features that a usable by people that aren't developers (PRs etc).
If you want a point to your side, GIMP is a bad hill to die on. Blender is a much better example of a tool that we should be teaching as opposed to Autodesk Maya (etc).
That's the thing with Blender though, it has feature parity with the things it's competing with and a community and network effect. For something open source to compete with GitHub it would need to be so much more than Git and offer the community and discovery that forms a big part of the "open" GitHub.
In the meantime people need to get things done. NOT teaching them the tool they're inevitably going to use is just a pointless battle to fight. Better to create the alternative first and lure people with straight up feature benefits.
> I notice from your blog that you have links to a number of commercial services including Keybase...
Blog has been under redesign-rebuild for a long time--other priorities. Said services were planning on being removed and replaced with alternatives as mentioned.
They always are the last to be updated (mine also).
We don't disagree in principle about open tools being better. I do however think that there isn't a viable alternative to GitHub at this point. The same could have been said about ExpertExchange at one point, and while StackOverflow isn't perfect, they do at make everything Creative Commons.
Tackling GitHub would require building a platform where all of the other tools such as Pull Requests, Actions, Issues were also represented in a portable format. Recognising that Git is just one interface to GitHub now rather than the main part of it.
Truth. Vendor lock-in is a concern when it comes to platforms. One thing that can help is writing shell scripts (or Nix building those scripts) for CIs to run minimally. With the merge request model having issues scaling, I'm curious how long til we push to a different model, or a different tool like darts/Pijul.
