RE: .il code/file questions
Robert Craig wrote:
<SNIP>
> Sorry for the slow response.
> This is a tricky question to answer.
>
> It may be premature to even try to settle this issue now.
> After all, 2.5 has been out for all of two days.
> Derek and Andy, you have little experience in modifying the
> front end, and neither of you has described what
> changes you would like to make. You might discover that
> you actually need back-end changes as well, to properly implement
> some feature. The Euphoria back-end was designed
> specifically to execute Euphoria programs, not a wide variety
> of assorted programming languages.
I think you underestimate us. It's quite obvious that we can't modify
the front end to make runtime changes like exposing the vtable.
We certainly can make changes like fixing the include syntax, or fixing
the namespacing system, or adding assignment on declare, or private
constants, or structures, or oop, or labelled sequence elements, or
making it possible to overload routine_id()...
> The Source Code product lets you make any changes you like,
> to either the front end, or the back-end. The only "catch"
> is that the interpreter you create must be for your own use.
> You can't distribute it to the world (unless it runs on a new
> platform).
How many people are going to spend the money and effort to build an
interpreter that only they can use?
> The Public Domain source lets anyone in the world, for free,
> make any changes they like to the way the Euphoria interpreter works.
> The only "catch" is that it runs slower than
> the official interpreter. Translating it helps, but it's
> still slower.
The PD source is not practical. I don't see how you can think it is.
We can make all the changes we want, great, but who is going to use it?
> I feel I'm already walking a tight rope, balancing openness
> with the need to protect my income. It should be obvious
> that I can't make it easy for people to produce their
> own "enhanced" versions of Euphoria that run at the same speed
> as mine, and can be distributed widely. That would be
> financially reckless, if not suicidal. I'm not saying
> this proposal would necessarily do that, but I feel the tight rope
> getting narrower.
Why would it be wreckless? You aren't selling the syntax anymore
anyways, you are selling the ability to make executables. How does that
change if we make our own front-end? We are still restricted to the IL
that the offical RDS runtime uses.
You should be worried about what will happen if you don't provide a way
for us to utilize the backend, because it will mean people making their
own IL runtime modelled after the very code you gave us.
Once someone creates an opensource backend that runs the official IL,
you will be in direct competition then. Not partnerships. Not offshoots.
You will definitely have to pick up your heels then to just stay afloat.
Sometimes the best way to keep things from being stolen, is to leave
them in the open.
> I also feel I have enough product configurations on enough platforms
> to keep me busy for now. Adding another configuration that
> would be used by only a few people wishing to create fast
> personalized versions of Euphoria would probably not be a
> good use of my development/testing/documenting time.
> But maybe I'll "see the light" after 2.5 has been out a while longer.
>
> Regards,
> Rob Craig
> Rapid Deployment Software
> http://www.RapidEuphoria.com
I think it's alot more than 'only a few people'.
Without the ability for us to take advantage of the optimized backend,
the PD source for the front-end is useless for making language
enhancements.
If we buy the source, we aren't allowed to distribute our changes.
So what's the point of having it available at all?
If you allow us a way to create IL code that will run on the backend, I
think you will find that the outcrop of new 'languages' will only
strengthen Euphoria's market. They will all rely on RDS's official
backend.
Every one of those offshoots is a new niche of potential programmers.
I don't see the threat if someone creates a new front-end for your
official runtime. If anybody wants the binder or shrouder, they still
have to register with RDS. Those offshoots are also advertisement for
the 'official' language. When those programmers that use an offshoot
that dies (which most fo them will), where do you think they are going
to go?
Chris Bensler
Code is Alchemy
|
Not Categorized, Please Help
|
|
