On Monday, 15 October 2012 at 22:14:48 UTC, H. S. Teoh wrote:
On Mon, Oct 15, 2012 at 11:52:19PM +0200, so wrote:
On Monday, 15 October 2012 at 21:29:11 UTC, H. S. Teoh wrote:

>If you think forbidding templates/STL is crazy, wait till you >hear >about the people who insist that const is evil and ban it >from their >codebase. (That was from before C++11, though, I don't know >what >their reaction would be now that key parts of the language >_require_
>const. Maybe they've migrated to VB or something. :-P)
>
>
>T

I can somewhat understand not using STL as the library assumes you are using certain paradigms, but they sometimes doesn't do the job. It is
similar to phobos being designed GC in mind.

But if you are not even using templates, why bother? For OOP? I would never move to C++ for OOP, since you are also losing something quite important in the process, interoperability with other languages. Most (maybe all) languages (i know) have some kind of interface to C. With
C++ you lose that one too.
[...]

I dunno, maybe they like struct names being specifiable without the struct keyword (I always find that awkward when switching back to C
after dealing with C++ code). :-P

It *is* a pretty crazy idea to prohibit STL, seeing as STL is what makes writing container-related C++ code bearable. I have horrible memories of the Bad Old Days when I must've reinvented linked lists at least 20
times, just because STL didn't exist in those days.

When templates first came out, I was elated that finally I didn't have to implement Yet Another Linked List. Perhaps it took that kind of experience to appreciate templates. :-) People who didn't have to suffer through these kinds of limitations often don't appreciate what templates
offer. (And that's C++ templates, with all their warts, not even
speaking about D templates -- which are on a whole 'nother level.)


T


You should talk with the Go guys which seem to be happy re-inventing the type of tools we used in C++, back in the days templates were still not available,
somewhere around 1993 in my case.

Does anyone remember the pre-processor hacks from Borland C++?

--
Paulo

Reply via email to