Replies

>And here’s where we come to the real issue—the whole reason that geeks can’t stand Internet Explorer:
>
>Geeks everywhere were forced to use Internet Explorer at work even when there are better browsers, forced to support it for corporate applications, forced to make sure web sites still work in IE, and we couldn’t convince everybody to switch to a better browser.

Why then do the geeks [especially the programming ones] not hate C, C++, and Unix?
C is fairly obvious, the lack of index-checking on arrays is the reason for most all of the stack-overwrite error/security vulnerabilities.
The typing system is horrible, characters in particular ARE NOT SIGNED! (And even if you were to define the "sign-bit" as a case-indicator that utterly falls apart for languages with more than 2-cases, like German which has a Title-Case).
The Syntax makes it easy to make errors, one of my 'favorites' is the assignment-in-condition-checks e.g. if (user = administrator) { ... };
Another is the the non-regularity of the operator-assignment shorthands; "!=" should apply 'not' and then assign the result.
The lack of strings is annoying and, in reality, the natural result of the lack of "real arrays" in C, that is they lack information on the valid indecies.
The OpenGL API is a good example of how nasty things the become in the C world; in particular because of the lack of procedure-overloading AND the lack of strong-typing, AND the lack of arrays.

As for OSes, Unix and its derivatives inherit a lot of the mentalities of C, simply because that language is so "close" to the OS.
The storage of everything, config-file wise, as [unstructured] text is fairly annoying. There are no guarantees on the structure of the text, as there would be in .INI files which retain the readability of plain-text while giving the advantage of grouping together 'attribute/value' pairs in its 'sections'.

Having said all that, it's odd that so many geeks embrace things like C/C++ after having been forced in academia to use C/C++, I would think that once in the "real-world" they would use better tools. Like Delphi (or even VB) for deving GUIs, LISP for lots of scripting tasks (EASY to write a parser), and so forth. The .NET ability to use multiple languages is probably underused with the MS pushing for everything/examples to be in C#; but I'm of the opinion that different languages for different domains is a good idea. Now, in order to avoid being wholly negative I should offer what I perceive as superior alternative to the C/C++ paradigm (imperative/procedural).
Ada:

has arrays with proper index-information,
low-level support (it was designed to write software for "non-standard" hardware such as missiles because of the DOD's commission of the project),
cross-platform portable code (Ada has an ISO specification, and was the first programming language to do so, all Ada compilers *must* comply with the ISO standard for Ada and pass a suit of validation tests),
has built into the language the TASK construct which allows the programmer to multi-thread AND have the compiler take advantage of multi=processor computers and further optimizations (it is widely regarded that the porting of code to take advantage of multiple processors is the biggest intimidate challenge to the industry),
Objects and java-style interfaces (Ada 2005),
'Input & 'Output attributes for Classes (allowing the reading-from/writing-to a stream; the 'Input attribute is needed for proper dispatching.),
'Read & 'Write attributes for objects themselves,
Generic support for: Packages, Procedures, and Functions,
has the Ada.Numerics package
has the Ada.Containers package,
has support for distributed-programming, again thanks to the TASK construct