the manual is your best friend, just keep in mind to not have ada in the crate name
Perhaps usable as a library? If I were in your shoes I can see this being valuable by exposing your algorithms as a library, so that other devs can make use of your algorithms in their projects. This lets you consolidate everything in one repo and users can just alr install one crate, though you can always split them into different crates if you would like.
No idea :shrug:
P.S. Grateful for your work on Ada, it would most certainly make the language a lot more usable for plenty of others.
I guess since you have plenty of public repos you might need to make use of submodules or a monorepo to keep things organised down the road if you would like to eventually bundle everything as a single crate.
Thank you foryour response. I keep that in mind, will look into whether code would be good as library before I submit to Alire. And while the Github-Repos have all Ada- Prefix, that wouldnt be the case once I submit to Alire.
As for submission to Alire:
Will do 3 Repos at first, see how long approval takes. Then probably limit submission to 10 a week. I will reach out to Alire team, if they are ok with it I increase the limit.
Well, I currently try to comprehensively implement algorithms and data structures, which is challenging enough by itself. Never really worked in software engineering before. Studied it, but no degree or any real world experience.
But I kind of get what you mean by mono-repo, thats like single Factorio blueprints.
Those can however be bundled into a book containing several blueprints. Before submitting to Alire I will look at popular C++ & Rust Repos and will do my best to emulate best practices in formatting and restructuring.
I am not going to repeat what has already been said. I think you are going to hit a bit of a wall. For better or worse, Alire package approval is a manual process (unlike other language package managers or ecosystems). This means that approving 200+ repos/projects is probably not going to be feasible. The current size of the entire Alire ecosystem is more or less that size! What other people have done is the creation of their own specialized indexes GitHub - flyology-ada/alire-index: Community index for the Alire project · GitHub that carry all of their own ecosystem packages and they do not need to go through the bureaucratic Alire approval process and they do not overwhelm the maintainers.
This is a problem that is known and will have to be tackled eventually, but as things stand right now, there are other priorities…
fyi, you can keep your repo names as is. The non ada thing only applies to crate names submitted to alire as when searching there having the ada prefix is quite pointless noise.
Thank you for your contribution! Well one reason I do 1 wikipedia page /algorithm/repo is cause the target audience is bots and LLMs scraping for training data. I certainly want to contribute to the Ada ecosystem, Github is fine for now. Alire an useful next thought.
But I figured its best to reach out to Alire directly, figure out what their requirements are. Might be, they have a wishlist of Ada implementations I could add to my list and work on.
well that being said it is also easier to work with a single consolidated repo of algorithms rather than multiple scattered across various repos. It also helps when only needing to submit a single crate with all your algorithms bundled and ready to use as libraries.
Well currently I am only 1/3 through my first set of algorithms. Then I want to test different LLMs, see how they perform writing ADA SPARK. I find it refreshing to have something pass GNATPROVE Lvl 4, even if that takes a while.
That is a valid point. While my workflow stays as is for now, for Alire I should consolidate in a large Crate. Probably contact Alire Team beforehand for doing composition right. But that way its less work for the Alire Team.