justsomeguy: the front end for Savannah has had one active developer for five years. If you can get people to help him with optimizing the user interface (including gitweb/cgit and other dependencies with their own upstreams) and translating it and its manuals, that would be nice for the project.
get people to help him
I wish I knew how…
F-Droid is dependent on microsoft github, and should be tagged as such.
Shoud that be considered an anti-feature?
Moved here from there: IMO, something should be done to encourage moving away from non-free, proprietary web dis-services for development. I am convinced there are so many apps on proprietary Github/lab, it would almost be pointless to tag them all. But, some way of distinguishing, for easy filtering, would be a good thing. Positive/Pro-tagging those apps on F/LOSS development services is something to consider.
F-Droid app particularly, because they should set the example, but remain on proprietary gitlab and depend on so many apps on proprietary github/lab… deserves the non-free net dependency tag, under current tagging.
I am a pragmatist, so in these circumstances I am in favor of working with what we currently have as we strive towards perfection. I think tagging things that are suboptimal is a good start, because it indicates to people that the status quo is not desirable.
I say this as a developer who has made the decision to make whatever sacrifices necessary to host all the critical infrastructure that my project depends upon (code repository, website, issue tracker, forum, email server, etc.). I do use a few communication resources that I don’t host myself (Fosstodon account, F-Droid’s forum), but none of them are central to my projects, and if they all went away the projects would continue to function without them. So, I have some sense of how much effort is required to only use open source tooling and to insist on self-hosting on top of that. And I get why many developers do not do so.
Self-hosting is probably a step beyond what is ideal for all projects, but I think it isn’t unreasonable for open source projects to strive towards a situation where all the SaaS they depend on is free software.
I understand and empathize. As someone with idealist tendencies, I feel sometimes “the only winning move is not to play the game.” (War Games) If we can’t make android apps without sacrificing principles, maybe it is better to work on something else like apps for Linux phones, or ways to make android apps without sacrificing principles.
Regarding Android itself, Google is moving it in the wrong direction (it is becoming less free with each release). I have spent a bit of time pondering over when it will cross the line that it becomes unusable as my daily driver. We haven’t reached that point yet, but I think it will happen at some time down the road.
I am pretty much against flagging usage of github as an anti-feature. Most of all because it is not a feature in the first place.
a bit longer:
Excuse me for lecturing but there is something that might not be fully understood by everybody:
- In git every commit has a hash (which is used as its identifier. So it’s not optional)
- The contents of the source code files (actually the change set) are a part of how that hash is computed
- Nobody can “sneak in” code without changing the hash (and changing the hash wouldn’t go undetected since developers would get crazy merge/pull conflicts all the time)
So let’s stop for a second an appreciate that due to how git is designed it will never have an effect on the actual code whether, at some point in its lifetime, it was hosted on proprietary git servers. Good job Mr Torvalds.
In other words: You will never find any undesired behavior in an app and later the analysis will say “Oh! This is because they hosted the code on github.”
If we can agree on that, the question arises why we would want to flag github as an anti-feature? Basically that would mean to tell users out there “There is something that will never in any way effect you or your loved ones but we kinda don’t like it so here’s an anti-feature flag”.
IMHO that is “spreading FUD” in the best case.
But I honestly think what’s even more harmful is that it waters down what anti-feature-flags (in my understanding) are actually about: informing users of actual threats and undesired behaviors in the app that they don’t have the time or knowledge to analyze themselves.
UX guys have a name (can’t recall it. I am a backend dev…) for useless warning messages that only teach you to ignore and “click away” ALL warning messages. I feel that a “hosted on github” anti-feature will have the same effect.
If somebody feels uncomfortable with the way some open source developers spend their free time on their FOSS projects that is ok (although I, for one and in this case, couldn’t care less). If somebody wants to fight Microsoft that is fine too.
But an anti-feature flag is not the right place and “barking up the wrong tree”. You could e.g. write to the developers directly.
If that anti-feature flag should ever be put in effect I also strongly suggest being consequent and making developers fill out a questionnaire whether their app was developed with a FOSS editor, in a free OS, running on open-source hardware.
One could say the same about Windows or Whatsapp or Google Play, right?
I’ll quote myself from elsewhere:
Better think of it like this:
- Saying you like FOSS yet use Github you keep saying to your contributors that: “you should use <closed source git service>, I’m just quirky and odd”
- Every day you are still a Github user you keep your contributors prisoners too!
- If you don’t stand up for your principles, why should your contributors?
Like RMS says, every “NO” helps: Saying No to unjust computing even once is help - GNU Project - Free Software Foundation
( Yes, I adapt the text for each situation )
Well no. One could not. At least not if one does not want to compare apples and oranges.
You are mentioning software that in itself is “tainted”. You could write WhatsApp only with the purest of FOSS tools and it would still be problematic.
But with github we are talking about how software is made and in this case it has zero effect on the endresult.
Yet you want to add an “end-user facing” flag? Makes no sense to me.
That’s as if books in the store had labels like “was written in a non-ergonomic workplace”. The concern is valid. But the audience is totally wrong.
You want to change how software is made? Address the developers. Not the users.
I like how you did not respond to the content my post and then left out the 3 words in that line that do NOT serve your point.
Read my postS from above again, what is my point again?
Your point is (to paraphrase Adorno) that there is no right life in the wrong one.
What was mine?
You can see my posts above that I try to explain that such an AF is unneeded and that F-Droid is not the place to annoy devs.
But I do agree that this is an unfortunate situation.
I think an apology is in order.
I completely read the thread about two months ago and only re-skimmed before posting my statement. When your reply came I must have pulled the wrong summary from my mental file cabinet.
Sorry. Now it was ME who’s barking up the wrong tree.
Guess I was already in “fight mode” because to be honest I find the whole issue rather annoying:
As I tried to explain there is no harm done on the source code level.
Also there is no real hard lock-in here. Pull the repo, push it somewhere else. Or keep it local. The beauty of a distributed version control system…
In other words: I find the discussion to be rather theoretical and philosophical (<- that does NOT mean “pointless”!) and the impact of code host in software development is quite over-exaggerated here.
And then I think of the DEVs. Good ladies and gents that (often after a hard days work) sit down and code something for all of us to use for free. I feel deeply thankful (also because I myself can’t really bring myself to code ANYTHING after getting home and putting the kids to bed . . . )
Planing to tell these people that their app will be put into the same category as apps that send data home to 3rd party servers just because somebody deems their toolchain unworthy feels a bit spoiled and bratty to me.
(and that kind of feedback could be the straw that breaks one or two or 20 camels backs out there. That’s not the price I would like to pay in the name of purity)
I am being a bit emotional here. Which is not always good in a discussion.
Does nobody care if microsoft github is profiting by creating new products like copyalot off the backs of developers who gave them their source (as well as data on interactions)?
Thanks for the info on how git works. In a perfect world, maybe FreedomBox would make it easy to self-host, and actually take advantage of distributed development instead of centralized.
app will be put into the same category as apps that send data home to 3rd party servers just because somebody deems their toolchain unworthy feels a bit spoiled and bratty to me.
“It is what it is.” There are 11 categories. Each is simply a line of words in app descriptions, not something to click to agree. OsmAnd has 4, no biggie to add one more for using proprietary microsoft github.
Wanting to use a more free toolchain seems like a desirable thing to me, and to some developers. I understand it is nearly impossible to maintain, however, because keeping up with Google’s changes may be harder than leaving proprietary github/lab.
I have some understanding of factors that push dev’s to use proprietary github/lab and others, from here and elsewhere.
informing users of actual threats and undesired behaviors in the app that they don’t have the time or knowledge to analyze themselves.
Again, a “positive” recognition, or Pro-Feature, could also be used to “reward” apps using “better” free development sites, to avoid “shaming”.
Yes, 777 apps with many anti-features: 220 with No Source Since, 25 with Non-Free Addons, 27 with Non-Free Assets, 35 with Non-Free Dependencies, and 380 with Non-Free Network Services.
IMO, the 3000+ F-Droid apps developed on Non-Free Github/lab do encourage using “Non-Free Network Services”, almost as much as the 380 that are currently tagged.
apples and oranges
I hesitate to use analogies because they often go off the rails, but lets risk it - Food labels. We could debate whether labels like “organic”, “fair trade”, “contains no high fructose corn syrup”, processed in a facility that also processes nuts, contains only free-range chicken, kosher, halal, vegetarian, vegan, … are well defined, meaningful, or just marketing BS. Regardless, many people want to know some details about the creation of the foods going into their pie holes. Similar applies to F-Droid and processing of apps to put into your android, or should.
You know what?
Technically I am one of those developers (I have some pet projects nobody needs and I fix a bug once in a while).
And I really don’t care.
And most people in the open source community that I interact with seem to be concerned more with “getting shit done” and less with “saving the world with rigorous principles”.
Looks like this is also reflected in the license statistics from 2015. Basically all projects allow commecial use. And more then half of them are released under licenses (MIT and Apache) that explicitly allow changing the code and then making it closed source…
Those developers don’t need saving, can speak for themselves and are gone from github within 30 minutes (longer if CI is involved) if the cost-benefit ratio does not work out for them anymore.
(a little anecdote around copilot: when the news broke last year I overheard two colleagues joking that if Microsoft wanted to build a tool that helps you write shitty code, they chose the right training set)
you don’t need freedombox for that. Just a linux machine and about 20 minutes.
But self-hosting comes with its own problems, issues and hassles. I am not doing it.
And while git in theory can be totally decentralized in real life projects are going to converge towards one dedicated upstream host.
“wanting” : yes
“being pushed” : not so much
Thank you for commenting. You have helped me solidify my views of apps and F-Droid. We simply disagree on a lot, but that’s OK.
People have suggested F-Droid should have app ratings, but that is difficult to implement without tracking users, and has other problems including generally pissing off developers. OTOH, F-Droid could have their own “curated” ratings, based on F-Droid values (if they can agree on them). Anti-Features is a start, but only in a negative way.
I would like to see a “curated” F-Droid numeric rating system that includes plus points for Non-permissive licenses, using “free” development systems, and websites that don’t block Tor, regular but not too high frequency updates, …
Of course I would also like F-Droid data to be accurate and up to date, but people can only do so much…
Funny, but if we don’t know how they weighted stuff when training their AI, we don’t know if the shitty code was included or not. Maybe they put more weight on higher quality “private” projects…
If you “don’t care” about copyright or licensing violation, then IDK what can be said.
A “freenes level” is something I could get behind. To me that’s a completely different thing than an anti-feature (which to many is warning and sometimes even causes apps to be filtered for which I still don’t see a rational justification in using github).
The score would be much more in the “gamification ballpark”
That is not what I said or meant.
I do care about copyright VIOLATIONS.
But on one hand I don’t care about the copyright of my code (which is why I, like so many, chose the MIT license for my projects).
On the other hand as far as I underdtood the discurse it is very debateable whether Microsoft violated licenses (even those like the GPL) simply because using the code to teach an AI what code looks like is not formaly covered by those licenses.
Drew devault agrees - do not use, and push back against, Microsoft github and copyalot (copilot). No surprise.
After moving from ms proprietary github to gitlab, PostmarketOS is considering moving to Sourcehut.