Re: Are there "compiler generators"?

On Tuesday, May 31, 2022 at 8:20:08 AM UTC-7, Hans-Peter Diettrich wrote:
> On 5/30/22 2:53 PM, Hans-Peter Diettrich wrote:


> > analysis or translation of computer languages to be on topic. -John]
> What are those "computer" languages? I'd prefer "formal" languages
> (Chomsky...) instead. E.g. Meta§ also was used for DNA analysis.


> [I was going to say artificial languages but I don't think we have anything
> useful to say about Esperanto. In practice it hasn't been very hard to keep
> discussions more or less on topic. -John]


Definitely interpreters and macro processors have been discussed,
though some might not call them compilers. Also text processors
like TeX.


I am wondering, though, about (human) language translators.
It seems that many use non-deterministic AI systems, and so are
fundamentally different from most of what is discussed here.


If you parse Esperanto with a Flex/Bison parser then it should
be fine here. If you parse Fortran with a deep neural net, then
maybe not.
[That's essentially what I've been thinking. In the 1950s and 1960s
there was a lot of work trying to do human language translation using
formal methods and it worked very poorly, sort of adequate for
translating technical manuals, not for anything else. The breakthrough
was when someone at Google realized that the Candian parliament's
Hansard, the transcript of debates, had high quality parallel
French/English translations going back a century and they fed it
into their machine learning systems. -John]

Return to the comp.compilers page.
Search the comp.compilers archives again.