Re: Conformance
- From: "cr88192" <cr88192@xxxxxxxxxxx>
- Date: Sat, 2 Feb 2008 00:40:35 +1000
"James Kuyper" <jameskuyper@xxxxxxxxxxx> wrote in message news:GIDoj.11122$O9.2849@xxxxxxxxxxx
jacob navia wrote:
...
<snip>
I do not have that money. Then I tried to propose my extensions
through the C++ group since the C group was disbanded because
(as everybody knows but me) C is dead and the group no longer
exists.
I presented that paper in Novermber, and after many calls I stopped
calling them. Nothing happens.
Hey - has it occurred to you that people just don't think your ideas have much merit? Whether or not they are right in that opinion, they have a right to possess it and to act on it. Just because you think an idea is good doesn't obligate anyone else to agree with you.
hmm...
now, in my case, I don't really expect people to agree with me...
nor, are my ideas, particularly even considered as something that should be standardized...
This was tried before, and turned out to be a "slippery slope" with the
target clientele better served by simply switching to C++.
Yes. C++ is the better C, and all development of C, therefore, is
completely useless.
As usual, you're misreading his comment, and I can't see how you could be doing so unless it were deliberate. What he said is that C++ is better than C for certain types of programming, and C is better than C++ for others. That is quite different from saying that C is useless.
Reducing the differences between the languages only serves to put the two languages into closer competition with each other for the same kinds of applications, and C will lose that competition by reason of being behind the curve in comparison with C++. Any plan for C's future that doesn't call for it's extinction must must emphasize moving in the direction of what C does better than C++, not things that C++ does better than C.
and, my case, I am off in an even more bizarre direction:
it is my hope to be able to use C like it were Python or Lua or something...
but, why not then Python or Lua?
because these languages introduce a few other, terribly painful, artifacts...
firstly, I would have several languages:
one for the statically compiled parts;
one for the dynamically compiled parts.
secondly: FFI horror.
thirdly: at least some of us, think non-C-style syntax is ugly...
as such, if I can write both the app and the scripts in (more or less) the same language, I have an advantage...
as such, part of the ideal is to have a compiler, where I can tack on new, presumably domain-specific, extensions, absent too much pain (nor, necessarily, assume that these extensions be somehow implementable within the language itself, as is generally assumed, for example, with C++).
as such, the compiler becomes a tool, where we can slap on a few extensions, and use it as a kind of domain specific language.
geom vectors are an example (relevant to my particular domain), but there are others.
for example, it is not too unreasonable to imagining tacking on compiler extensions for database access, or XML processing, ... as well.
the code involved may be almost entirely non-portable, but it will also be, scripts...
I can make an analogy to English:
English is, in itself, incapable of self-extension. we can't write a few statements, and suddenly define a change in terms of word-level processing, vocabulary, or grammar.
none the less, there are many domain-specific variants of English (including specialized grammar, vocabulary, ...).
these variants are understood within this domain, but elsewhere, are essentially meaningless (there being a need for "plain english", which is often little if at all capable of expressing the ideas of these domain-specific variations).
similar, is not implausible WRT programming languages, and is often done, but not much within the C or C++ communities (which tend to be more focused on a "one size fits all" approach, and imagine the whole world as essentially systems programming, apps programming, or embedded systems).
however, standardization is good.
we don't need new and completely unrelated languages for each of these domains (this has often been done in the past, as, once one learns how, writing a new specialized interpreter or similar is hardly much work), as unrelated languages lead to unecessary fragmentation.
take, for example, one person does a distributed protein-folding simulation in a specialized Scheme dialect, and another instead uses Forth. users of one system, have hardly an easy time migrating to the other.
going back to the English analogy:
even though each domain uses a specialized variant of English, at the core, English is still English.
as such, a person who knows one domain can, absent too much additional work, learn to read literature from another.
for example, I am a programmer, but I can also go and, at my leisure, read information on biomedical research, and other topics.
if my parent language were English, programming being talked about using a tweaked out version of German, the medical field used Japanese, and my religion used Greek or Latin, I would hardly have an easy time here...
then again, no effort on my part is likely to defeat the prevelence of domain-specific languages.
<snip, rest of post...>
.
- References:
- Re: Conformance
- From: James Kuyper
- Re: Conformance
- Prev by Date: Re: Conformance
- Next by Date: Updated version of "The New C Standard" book available
- Previous by thread: Re: Conformance
- Next by thread: Re: Conformance
- Index(es):
Relevant Pages
|
