Does anyone else here positively enjoy programming in SQL?
I love working with it, it's my favourite. Especially T-SQL.
Name:
Anonymous2006-02-04 12:04
I like writing SQL, but I hate writing SQL within another programming language. Concatenating and escaping a bunch of strings to turn my language of choice into SQL seems pretty stupid. Not that I have a solution.
Name:
Anonymous2006-02-04 13:27
SQL is a query language, not a programming language. It's not turing-complete (the standard SQL is not, nonstandard extensions may be).
I don't like SQL. It's very easy to read "look, almost like English, lol!!", but then it's not "almost like English" when you're writing it, and you have to remember how to write such verbose syntax properly. That said, there's nothing better to query relational databases in a standard way, and I don't like data objects.
I just can't respect COBOL. Any language with the "Business" word somewhere means it's shitty stuff made to convince managers that it'll allow their company to maximize their profits, shorten software development cycles, and increase productivity. Management bullshit, to put it simply.
Name:
Anonymous2006-02-05 17:54
A lot of the things going into business targeted languages nowadays are the exact same things that any large scale software project needs. There are entire fields now where prima donna languages like LISP with custom macros and shit just can't cut it when you need many people of varying skills and abilities all working together to get it done in any reasonable time frame. If you are just a useless academic playing with algorithms, though, then avoiding the things business goes for is fine.
Name:
Anonymous2006-02-05 19:30
>>10
I've heard some arguments that LISP scales better than any other language. The main reason is because you can extend the language itself. If you're working on a massive scale, libraries aren't enough.
I don't know if this is true or not, but some people in Amazon were bitching for a while that Java, C++, perl, and whatever else they were using just didn't cut it for the scale they were working on. The Java frameworks most people think of when we say enterprise couldn't take it. They had to write their own frameworks in order to scale.
The two main contenders as a replacement for all that were Erland and Lisp. Until someone goes out and does it (Amazon might be first?), we won't know for sure, but it'll be interesting to see.
Besides, Amazon seems obsessed with competing against Google for the best and brightest. Those kinds of people already know, or are prepared to learn, such languages.
Name:
Anonymous2006-02-05 19:31
BTW, the "I don't know if this is true or not" comment refers to LISP scaling.
Name:
Anonymous2006-02-06 3:14
Until someone goes out and does it
Yahoo Stores was originally written is Lisp, back when it was Viaweb. They didn't tell anyone at the time, because they considered it to be their secret weapon; it enabled them to code features 10x faster than the competition.
After Yahoo acquired it, it eventually got rewritten in C++ because the maintenance crew found Lisp syntax to be scary. Of course since the software relied heavily on some of the most powerful parts of Common Lisp, not least macros, this meant that Greenspun's 10th law came into effect:
Any sufficiently complicated C or Fortran program contains an ad hoc, informally-specified, bug-ridden, slow implementation of half of Common Lisp.
There's a lesson in this. The lesson is that C++ sucks.
Name:
Anonymous2006-02-06 3:57
After Yahoo acquired it, it eventually got rewritten in C++ because the maintenance crew found Lisp syntax to be scary.
Apparently something similar happened at Amazon too. For a while, very early on, they had a sizable base of Lisp code, but then they hired too many people who didn't understand it. Guess what happened?
It's a bit pathetic, but it seems that many (most?) developers are just too lazy to learn a new language, so they try to force through what they're familiar with. The end result was C++ and Java (and a bunch of rusty developers).
>>10
Yes, because business use the tools they really need. They really increased productivity and performance with Java. Sun said so.
I can think of at least three languages more productive than Java, and at least three languages faster than Java, which would scale equally as good or equally as bad as it. In fact, Java's massive complexity and overengineering is a hindrance for scaling, not an advantage as Sun tells stupid managers.