BD Software Discussion Board
« No concept in C++0x »

Welcome Guest. Please Login or Register.
Dec 22, 2009, 7:32am

[Home] [Help] [Search] [Login] [Register]



BD Software Discussion Board :: General :: STLFilt :: No concept in C++0x
   [Search This Thread][Reply] [Send Topic To Friend] [Print]
AuthorTopic: No concept in C++0x (Read 121 times)
Happy VC6 User
Guest
No concept in C++0x
« Thread Started on Jul 24, 2009, 9:51am »		[Quote]
Stroustrup shares committee decision to drop concepts on DDJ: http://www.ddj.com/cpp/218600111

Is this good or bad news for STLFilt? Anyone used conceptgcc and been impressed with its diagnostics?
Link to Post - Back to Top  IP: Logged
Leor Zolman
Administrator
*****
member is offline

[avatar]


[homepage]

Joined: Jul 2004
Gender: Male
Posts: 94
Re: No concept in C++0x
« Reply #1 on Jul 25, 2009, 12:03pm »		[Quote]
Yeah, that was quite a shock to me. Herb Sutter doesn't seem to think it's such a big deal:

http://herbsutter.wordpress.com/

Personally, I'm worried that the eggregious template-related error messages are going to keep C++ away from the front lines at educational institutions...when undergrad CS students start out, would you want them dealing with those kinds of messages? I sure wouldn't. So they start with Java and .NET or something, which aren't necessarily bad choices, but there's perfectly good curriculum based on C++ (such as Bjarne's new textbook, which I helped him proofread) that is going to be passed over IMO because of the error message issue.

Hopefully the concepts people can keep grinding out implementations and we get some sort of de-facto standard, the way hash-based containers were implemented universally before they even made it into TR1, and then EVENTUALLY the committee bites the bullet. My prediction is that the next interim (5-year) update, due around 2015, will include concepts.

I did play with concept g++ a bit, and I like the overall approach of concepts, but it's not like I did any production work with it. For one thing, the test compiler is slow as molasses. But that's not due to any inherent inefficiency wrt concepts, just a matter of effort not having been put into optimizing the compiler itself.

In the meantime, I've had to step back from STLFilt evoultion/development in order to earn a living, and it's not clear how much energy I'm going to have to put into it any more. Especially when no one ever communicates with me about it any more. I'm not sure what interest there really is in having it evolve...
-leor
« Last Edit: Jul 25, 2009, 12:06pm by Leor Zolman »Link to Post - Back to Top  IP: Logged
   [Search This Thread][Reply] [Send Topic To Friend] [Print]

Click Here To Make This Board Ad-Free


This Board Hosted For FREE By ProBoards
Get Your Own Free Message Boards & Free Forums!