Raw Consciousness

SOFTWARE DEVELOPMENT - Effective C++ 3rd edition....

Hey, Effective C++ 3rd edition is out and it has a whole chapter on template meta programming....

If someone knows of a cheaper price than bookpool, let me know!

SOFTWARE DEVELOPMENT - C standard book and more...

Just found out through The Inquirer that Derek M. Jones (a C language guru from the UK) published, online, a free 1600+ pages book on the C language standard including tons of information from psychological/human aspects for code organization, to performance of different constructs, to differences in the current standard to C++ C compatability to a lot of other ton of stuff. 

I am particularly most interested in the psychological/mind issues, but there seems to be a lot of everything.

Here's The Inquirer link for perspective.

Here's the PDF downloadable link.

SOFT DEV - What's wrong with C++?

Bartosz Milewsk at Reliable Software has made a short but very interesting article that I think summarizes extremely well how I have felt my whole journey with the C++ language has been and where I see felt it's all ending up to be like:

...But if a language is disproportionately complex, it starts adding complexity to the problem you're trying to solve instead of reducing it. There are endless examples of unnecessary complexity in C++.

Accidentally, the parsing difficulties of C++ might be the biggest roadblock in its evolution. ...As long as C++ gurus live in the clouds of the Olympus, they won't see the need for this kind of evolution.  That's why C++ becomes more and more elitist. In the future, people who do NYT crossword puzzles and the ones who program in C++ will be in the same category.

and...

... C++ is becoming a freak language that's parading its disfigurements in front of mildly disgusted but curiously fascinated audience.

"So you have to put a space between angle brackets? How bizarre!"
"Are you telling me that you can't pass an instance of a locally defined class to an STL algorithm? How curious!" ...

I won't quote more. Read the article, it's very amusing, interesting, and ummm... true. At least to me it is.  I love the point he makes of C++ and Java/C# in all this. How true.

I have felt at times that those who cherish their long-hours earned skill in C++ sometimes defend their L33tz programming skills sneering at the "lesser programmers" the same way someone who loves Linux a bit too much sneers at the concept that pershaps some tasks can indeed benefit from a more GUI/easier interface approach.

I believe Joel called this "those who sneer at those programmers who code in a language that doesn't have blocks of { }'s."

SOFT DEV - MS Managed .NETness...

I really don't have much experience with the .NET framework (yes, I think I'll dive in now), but I have done some cursory looks at code. To me the code looks way better that what MFC or plain vanilla win 32 are supposed to do -  of course, credit where credit is due, C# is highly influenced by Java. at The expense and price seems to be some loss of control (or apparently so) in how memory is managed.

It occured to me that for many applications on Windows, nodoby really cares about the later.  I played a bit with the latest Visual C# Express (aka Visual C# 2005 Standard Edition?) and there are some auto complete enchancements, refactoring aids and debugging aids that are just too cool.

So the real problem in my opinion is the .NET framework. The latest beta I think weights at 12x megabytes? (probably shorter in the release form).  That's way too much for anybody to download.

It occured to me that there could be a solution to this, but it will require Microsoft to invest in "making sure .NET is spread all over."  If Microsoft could strike a deal with Best Buy, Circuit City, Frys, Compusa that gives away say, the word processor inside the latest Works, or some free photo editing on CD - along with the .NET framework- this may at least help immensely to spread the .NET platform.

I find interesting that the very latest ATI video drivers do require the .NET framework installed to open its new n' shiny (tm) control panel.  So all hardcore games with LeetZ Ati cards will probably have the .NET framework installed, but that's not average Joe or Mary.  So if Microsoft wants this platform to be spread all over, I think it will have to give away something for free but that requires the .NET framework installation.

Otherwise, I honestly don't see .NET taking off for the next five years.  And five years is a very long time.

SOFT DEV - interesting & dangerous...

Just when it was mentioned that structs can't inherit from others recently I saw some code to that end. But the interesting part was that the programmer in quesiton needed to "memset to zero" the whole thing quickly, it having a bunch of fields BUT it also has virtual functions (a struct! :-) ).

Anyway he told me I would probably be able to appreciate (in a laughing way) what he did.. and actually I did. It goes something like this:

Blah:ResetBlah()
{
memset(this, 0x00, sizeof(Blah)); // may be not 100% like the code was but it was a one liner
new (this) Blah(); // a placement new unto itself!
}

So you know, if you have virtual functions, you have some vptrs... so memset to zero will wipeout not only your fields but also your virtual table ptrs... but the placement new takes care of replacing them. And hey it does work. I wasn't scared of this code once I realized some months back that placement new is a "runtime dynamic C++ linker" of sorts.

I did warn him though, that if he had members that were tracking some resource or something, a memset zero obviously sends all that tracking to hell. So we both laughed as we reached the verbal conclusion: just add a call to destruct itself before the memset.

Hmmm... what can I say. I just found it interesting and dangerous... and also very funny.

SOFT DEV - C++ huh?!

Well, new category. The geek in me had to come out of the closet I guess...

I was experimenting with some design stuff (doing some prototyping) an isolated something that caught my eye in C++:

class A
{
public:
float x,y;
};

struct tTest
{
A &r;
A *p;
};

A myA;

tTest myTest = { myA, &myA };

myTest.r.x = 33;

Huh? I didn't know I could do that... moreover, I looked at the memory and the pointer and the ref are identical. Hmm.. hell Raist, what did you expect.. magic? Remember C++ is OO assembly. I must have misunderstood Vince a while back sometime, I thought they were something else for some reason (the references).