Could LLMs help here to translate between programming languages? Same with COBOL. Is a coding assistant the way to get legacy systems updated -- sounds like a consulting niche.
For Fortran, it is really a niche. This is a part of what I am doing for a living. For the first time, I proposed my service on the traditional "who is hiring/looking for work/freelancer" to see if some people reacted to it because of a previous discussion about this report[0].
For the moment nothing, I am not really surprised. It is quite a lot of you know someone who... kind of niche. To port or upgrade some scientific code from Fortran to a new language, you need to have some minimal, some times extensive, domain knowledge, Fortran knowledge and target language knowledge.
It is hard to find the right persons, with not that many persons, you do not want to start a big migration and lose the head developer on the way, etc. This is why if you read in the report and the forums, what people want is to "reboot" the community/ecosystem and put new life in it. But this is also hard if the ecosystem starts to fossilize.
I love and really enjoy working in Fortran, but as a paradox, because of the mentioned situation, it is not that easy to find Fortran work.
I doubt it. Non-idiomatic translation is already covered by compiler-based technologies that already exist and are off-the-shelf. Idiomatic translation would require an LLM that could ingest large chunks of the code base; it is not sufficient to translate one function at a time (which is, again, a solved problem), and LLMs have scaling problems still as you go up. And, this is code, not human language. It doesn't take very many small mistakes where the code did this one obscure thing, but the LLM thought it was a more popular thing, before the whole translation is worthless. In my playing with ChatGPT if you deliberately bait it with code that looks like some really popular task but it's actually doing another, it is very strongly attracted to the popular thing.
An AI that uses LLMs but is more than just an LLM, maybe.
This is the right answer. Almost always in manual translation processes, there are questions about intent, either ambiguities in the original source (is this a big or do we keep it?) or clarifications in project planning (this code is unused... should we leave it out?).
