I am an academic researcher working primarily in high-performance scientific computing, with a focus on numerical methods and large-scale simulations (HPC environments, CPU/GPU clusters).
I would like to ask whether there is currently an active Slack workspace associated with the Fortran-lang / Modern Fortran community. I have seen references to a Slack channel in older discussions, but I have not been able to locate a valid invitation link.
If the Slack workspace is still active, I would appreciate any guidance on how to request an invitation.
Additionally, I would like to ask about the current role of Slack within the community: for HPC-oriented users, is Slack still used for technical discussion and quick exchanges, or has most interaction effectively migrated to Discourse and GitHub?
Thank you very much for your time and for your continued work on supporting Modern Fortran in scientific and HPC contexts.
For HPC context not Fortran specific go to HPCSocial https://hpc.social/ very cool community.
For Fortran specific I think the discourse is probably the best place. Moving conversation away from here could be detrimental to the community. I’d say that if you want to chat with someone you can use the chat function here and/or creating group messages.
I don’t know which Slack you are referring to, but there was a Slack channel for GFortran developers which has been migrated to a Mattermost (a free alternative to Slack) instance a few years ago, and is managed by @JerryD .
I agree with @jorgeg in his encouragement to use this Discourse, because:
It’s been my experience that activity for more specialised topics quickly stagnates when it’s scatted across different communities and technologies (Slack, Mattermost, Matrix, etc.), unless it’s centred around a specific project that many are involved in (e.g., the development of a compiler).
There are already a lot of “scientific computing” people here, and at least a few who worked or are working on HPC systems (@jorgeg and myself, for example). They may be interested in the discussions but unwilling to join (yet) another community.
Others who are not working on HPC systems may still be interested and able to contribute to discussions with their Fortran expertise here.
There’s the option of creating chat groups if you feel something is too sensitive to discuss in public. However, I would always encourage a more open discussion, as it may be very helpful to others in the long run and you’ll get input from more people.
If scientific computing/HPC ends up as a very frequently discussed topic, there’s alway the option of creating a new category in this Discourse to stay organised.
There is https://society-rse.org/ and they have a slack presence and a Fortran channel. Here is some blurb from their site. “The Society of Research Software Engineering was founded on the belief that a world which relies on software must recognise the people who develop it. Our mission is to establish a research environment that recognises the vital role of software in research. We work to increase software skills across everyone in research, to promote collaboration between researchers and software experts, and to support the creation of an academic career path for Research Software Engineers.”
Though not Slack, Reddit/HPC might also be useful for some topics (though it may be for more casual conversations…) www.reddit.com/r/HPC
I think it would also be great if Fortran Discourse would have sub-categories about general scientific/technical computing and HPC (preferably related to Fortran, but with other topics also acceptable to some extent).
I think that topics can be added, it might make sense to create topics like “Scientific Fortran” (even if it might sound like a pleonasm) or “Scientific Computing” or “HPC” ?
“scientific computing” and “HPC” may not be necessarily the same in topics (?), so I guess each category name may be useful. But first of all, I am not sure what “scientific computing” exactly means (though I think one can use such a tag if one feels that the topic belongs to that category.) In my case, I am most interested in computational algorithms (particularly modern ones) and related codes etc for such a category. Even “which random number generator is good in 2026?” is interesting to me
FYI, I tried using the search bar on the Discourse page with “scientific computing”. It gives > 50 hits, but some of them are “false positive”, so an explicit tag may be convenient to look at related threads.
That is an interesting question. I think there might be a few topics that involve high performance computing that are not scientific. Mathematical proofs might be one example, say, for example proving the four-color theorem. Predicting stock market trends or some other economic behavior might be another example. However, even in these cases, the underlying tools, the mathematical libraries and the communication libraries, are the same tools that scientific computing uses. I watched a segment on my local PBS news yesterday about reproducing artwork with computers. The idea is that physical art was not designed to last forever, so if famous paintings can be preserved digitally, that digital information will outlast the actual physical art itself. They can scan details down as small as about 1/10 of a human hair width, preserving information such as brush strokes and paint texture. I thought that if computer scans are that accurate, then how is it possible to detect currency counterfeiting. Modern computing got it start during WWII with code breaking algorithms. That continues today with encrypting/decrypting encoded messages for bank transfers, cryptocurrency, military intelligence, and so on. Is that scientific computing or not?
Whether or not something is science is quite clear cut (it either uses the scientific method or not). If computing is done to address problems in science, it’s scientific computing. You can “do science” in the disciplines you mention, but they’re not scientific disciplines. So computing done in these disciplines can be scientific computing, but doesn’t have to be (even if the same algorithms are used).
Therefore, a broader category would be needed to include these (and engineering) if it is to be an accurate label. On the other hand, it may be more practical to use a description that doesn’t fit perfectly, as long as users see the relevance of that category to what they are interested in.
Indeed. They are very different. There’s non-scientific work done on HPCs, and the majority of scientific computing does not happen on HPC systems. (Even when it would make sense to do so, access may be limited. HPC system maintenance is very costly.)
I find that Discord is a great platform for things like this (and much better than slack).
Discord is free to set up a server and it could be a great place to have a Fortran server.
However, I also agree that spreading out a small topic like Fortran to too many platforms can be an issue.
Perhaps a Fortran discord server could be a place geared more for developer support (compile flags, syntax, stdlib, etc.) and this discourse more about the language itself (such as future additions to the standard, etc.). Both Intel and NVIDIA have discord developers servers which seem to be pretty active.
Or, perhaps discord could be for event announcements and discussion (such as Fortran talks at meetings)?
(though I think one can use such a tag if one feels that the topic belongs to that category.) In my case, I am most interested in computational algorithms (particularly modern ones) and related codes etc for such a category. Even “which random number generator is good in 2026?” is interesting to me 