[Oberon] What is the status of Lola-2 and its use intheFPGAversion ofProject Oberon?
paulreed at paddedcell.com
Fri Mar 15 15:11:43 CET 2019
> some parts [of SystemVerilog] are
> really useful, e.g. a subset of the assertions which even allow formal
> verification. That's something I personally would miss in Lola-2.
Or you could add them, in the same way that people have added features
to Prof. Wirth's programming language compilers. But for a whole host
of reasons, take care to differentiate.
> But from your response I conclude that Lola-2 is sufficiently
> expressive for a majority of real-world designs with the exception of
> some vendor specific features, but these can be encapsulated into
> something (Verilog or whatever) with a Lola-2 compatible interface, so
> a Lola-2 developer would not have to bother with their implementation
> details. Plese correct me, if I misinterpreted you.
No that seems a fair interpretation of what I've said.
>> I've experimented with doing some of that, in various ways.
> Is this publicly available somewhere?
No. Just idle experimentation. I also fixed a few bugs in arachne-pnr
in passing, and that wasn't hard either (I now use nextpnr). So I would
say the two sides together form an ideal base for getting into this
further, since all the source is available (admittedly in several
different languages, but that shouldn't be a barrier to someone with a
reasonable amount of experience and a sensible worldview).
One example of where Lola was nice for me, was that I very quickly
changed the backend to add code to the Verilog to aid simulation (e.g.
initial value setup for registers) without changing the Lola source code
> my point was that it makes little sense to use Lola-2 when
> the developer has to fall back to Verilog constantly because only a
> fraction of the design can be expressed in Lola-2 or for testing and
> simulation purpose.
I agree, if (rather than when) that were the case; I don't believe it
is, in general, and certainly not in my experience. But of course it
all depends on exactly what one is trying to do. Without specifics,
obviously I can't really advise!
Also Nemo thanks, I don't think it's at all irrelevant to add how these
languages originate, it can often be extremely helpful in judging how
things are going to be "under the covers". Especially when it's
usually, inevitably, bad. Or designed by committee. Or both.
All very relevant to people who believe Prof. Wirth has a point or two.
More information about the Oberon