Re: [SharpOS Developers] MOSA / AOT

[Bruce M. <ill...@gm...>]

> The MOSA forums are not used much right now. I agree with you
> there. You see I'm in a weird position - I'm contributing things
> and trying to drive the effort my way. I've been primarily doing
> things by communicating with those really interested directly in
> person and have used the resources SharpOS and Ensemble have provided
> me with. The infrastructure (like Phil said) is being worked on and
> will appear sooner or later.
>

And don't think we weren't keeping that in mind when we put you on the
SharpOS board. We are here for community, and we don't want to alienate
anyone. (i.e., I wasn't trying to alienate anyone.) Just keep the
communcation channels open. That is the key to us all uniting under a MOSA
banner.


>
> > Calling it a MOSA compiler may not be accurate if it has to
> > be branched four ways from Sunday in order to work with each
> > project.
>
> No branches. The design and architecture of the jit compiler is
> flexible enough to survive differing OS designs, calling conventions
> etc. I don't want to call it a piece of art because its not there
> yet everywhere, but it will eventually.


I hope you haven't bitten off more than you can chew. This is certainly
something to look forward to!


> > MOSA needs a domain. MOSA needs a mailing list. MOSA needs
> > *people*. Then, I think, MOSA can sit down to create
> > standards. There is so much to talk about, that I think we
> > need some clear, defined space to do it in, so we can track
> > what is going on.
>
> I fully agree with you here. Like Phil said, we're working on it. By we I'm
> talking about Adam, Phil, Scott, rootnode and myself. (Sorry if I forgot
> someone,
> as always I'm on the run.) These are five people with differing interests
> and
> focus coming from various backgrounds that are actively contributing in
> various
> ways there. The space is being worked on, however what hasn't been
> discussed
> yet
> is where we'll start. More on this below.


I might have displayed more interest in MOSA from the start (I think I did
sign up for a membership on those forums right when they were first created)
if it had been a bit more organized, and I know Will was interested - but
this isn't the first time we've talked about "getting along" with the other
C# OS projects, and we still really haven't proven that it can be done on a
meaningful level.


>
> > When you guys first started talking about MOSA, I didn't
> > think about code at all. But code is an understandable
> > extension of agreed-upon specifications. Specifications that
> > we need to take time, space, and conversation to figure out.
> > The MOSA AOT is fascinating and exciting, but conversations
> > about it are moot for now.
>
> This is the point where my opinion differs by far from yours. I consider a
> compiler and runtime environment *the* two *key* pieces of any effort.


I'll stop you right there. At that point of my ramblings, I was more
referring to my first impressions of what the idea meant. You've already
shown me the light on this regard.



> If
> your
> compiler isn't capable of certain things, then you can't use them. Which
> means
> that you either start adapting your designs to a lacking compiler (like
> SharpOS
> is right now) or you freeze those efforts and define a specific set of
> features you
> need from the compiler in order to do a *good* design. The latter approach
> is
> IMO the right one and the one I'm pursuing.


But as your compiler matures, people will start tinkering with it. And if it
has limitations, chances are people will code around it. With no documented
standards, we can't even write unit tests to help you out. I want to see
this compiler soar - but its going to take some bearucratic processes to
pave the way. Otherwise, it isn't any different than the path the SharpOS'
AOT took. (A lone developer dropping a black box on hungry
kernel-coders-to-be.)


> And please stop talking about the
> compiler as an AOT - in fact it is not.


Oh goodness. I didn't mean it as an insult - its just from habit. We all
spent alot of time here with our AOT - and we've seen other projects try the
"Its a magical box that compiles your code to anything" thing, and that
project ended up re-prioritizing (for that time being.) Its no small feat -
and I don't envy the tasks you have in front of you.


> And one of my next contributions
> will add
> jit trampolines


I like trampolines. They make me happy!

I know you won't agree with me on some of these thoughts and this is really
> ok.


I think we agree more than I let on. But since I don't have the time to help
you code, (which I really wish I could do), I'm at least taking the time to
suggest how this needs to be handled administratively.