search

rems-project / sail

Sail architecture definition language
Other
617 stars 112 forks source link

Plans for scaling Sail to widespread usage? #112

Open martinberger opened 4 years ago

martinberger commented 4 years ago

Sail has been soaring recently, congratulations!

In parts because of the association with RISC-V, Sail is getting a lot of attention. I've been in RISC-V groups where the lack of Sail skills is a major problem. How is the Sail community planning to scale to accommodate industrial uptake?

Right now, Sail is managed on the side by academics. This is perfect for a research project. But if Sail continues to see more usage, which I hope will be the case, academics will be overwhelmed with support queries, pull requests etc. The more successful Sail becomes, the more people without a background in programming languages, theorem proving, dependent types etc will try to use it, and ask basic questions. I wonder what the Sail community's plan is for scaling to accommodate its success? Sail foundation, Sail startup? EPSRC funding, industry funding etc.

I think it would be good for the Sail community to outline how this scaling problem is to be tackled. What's your roadmap?

Alasdair commented 4 years ago

Sorry if it seems like we've been ignoring this issue, it's just a big topic for a github issue! I'll make sure to bring this up this week at our weekly meeting

PeterSewell commented 4 years ago

On Thu, 29 Oct 2020 at 10:56, Martin Berger notifications@github.com wrote:

Sail has been soaring recently, congratulation!

In parts because of the association with RISC-V, Sail is getting a lot of attention. I've been in RISC-V groups where the lack of Sail skills is a major problem. How is the Sail community planning to scale to accommodate industrial uptake?

Right now, Sail is managed on the side by academics. This is perfect for a research project. But if Sail continues to see more usage, which I hope will be the case, academics will be overwhelmed with support queries, pull requests etc. The more successful Sail becomes, the more people without a background in programming languages, theorem proving, dependent types etc will try to use it, and ask basic questions. I wonder what the Sail community's plan is for scaling to accommodate its success? Sail foundation, Sail startup? EPSRC funding, industry funding etc.

I think it would be good for the Sail community to outline how this scaling problem is to be tackled. What's your roadmap?

It's worth noting that, even if Sail becomes very successful (as we hope), the community scaling problem is quite different from that of a programming language: ISA definitions end up with very many readers but rather few writers, and we wouldn't expect that many different Sail definitions.

The language also seems to be becoming relatively stable. We've been through a couple of major versions, and there's ongoing work on new tooling, but the language itself hasn't changed much for a while.

For new community members, there's already a bunch of documentation and some on-line talks - we'll see as usage grows what else is needed. Input welcome on that, obviously.

best, Peter

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/rems-project/sail/issues/112, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABFMZZTNXIGU2ELBJZSVBGLSNFC6ZANCNFSM4TDSB4GA .

martinberger commented 4 years ago

That's a good point, Sail is a mostly-read language, unlike normal PLs, which are mostly-write. History of ADLs portends a rapid sequence of enhancements, once a language becomes popular. (E.g. Verilog)