Language:
switch to room list switch to menu My folders
Go to page: First ... 66 67 68 69 [70] 71 72 73 74 ... Last
[#] Wed Dec 08 2010 19:07:16 EST from dothebart @ Uncensored

Subject: Re:

[Reply] [ReplyQuoted] [Headers] [Print]

c just says, long >= int >= word....



[#] Wed Dec 08 2010 22:42:53 EST from Ford II @ Uncensored

Subject: Re:

[Reply] [ReplyQuoted] [Headers] [Print]

yes and I always thought that was rather stupid. Kinda like objective C, oh if we support it, we'll handle it, if we don't we won't.
Real handy for the guy trying to write the program.

Anyway. I was just a little surprised that even pointers were 32 bit, but I guess it makes sense, I've just been spoiled by big hardare for the past bunch of years I forgot how it used to be.

[#] Wed Dec 08 2010 23:49:24 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

you all know I'm not a big fan of autoconf and automake and all that and I just realized another reason why it's not great: crosscompiling.
it only works if you're autoconfing on the machine you're building on and building for.
Maybe I'm the only one stupid enough to try and get joe to build and run on my g1, but the makefile and autoconf.h that it made were largely wrong and useless.

[#] Thu Dec 09 2010 06:12:21 EST from dothebart @ Uncensored

Subject: Re:

[Reply] [ReplyQuoted] [Headers] [Print]

you probably should have a look at what cirrus does with xscoar.



[#] Wed Dec 15 2010 21:29:47 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

http://adtmag.com/articles/2010/12/15/apache-quits-jcp.aspx


Frankly, I think I side with oracle on this one, not that I think oracle is right, but I do so dislike what apache is doing more and more.

[#] Mon Dec 20 2010 14:20:16 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Java has clearly jumped the shark with annotations.
I'd go on a rant about it, but you already know what I'm going to say.
I'm just surprised somebody hasn't written a recursive fibonacci figured-outer with annotations rather like has been done with the C precompiler.

[#] Mon Dec 20 2010 18:48:17 EST from dothebart @ Uncensored

Subject: Re:

[Reply] [ReplyQuoted] [Headers] [Print]

 

Mo Dez 20 2010 14:20:16 EST von Ford II @ Uncensored
Java has clearly jumped the shark with annotations.
I'd go on a rant about it, but you already know what I'm going to say.
I'm just surprised somebody hasn't written a recursive fibonacci figured-outer with annotations rather like has been done with the C precompiler.

they call it code generators these days.

google for "don't repeat yourself" aka dry



[#] Thu Dec 30 2010 18:56:27 EST from LoanShark @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Java has clearly jumped the shark with annotations.

Spoken like a man who's never built a custom annotation with Spring AOP.

*runs*

[#] Fri Dec 31 2010 08:19:55 EST from LoanShark @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]


Purely Functional Data Structures (some guy's PhD thesis, PDF)

http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.62.505

[#] Sat Jan 01 2011 08:03:00 EST from dothebart @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

interesting...  the ruby community also starts to drift into the direction of http://nodejs.org/ :

http://news.ycombinator.com/item?id=1533243

(i'm currently plugging some of nodes foundation libraries into citserver; libev and c-ares...)



[#] Sat Jan 01 2011 10:55:03 EST from IGnatius T Foobar @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

I wonder how long it'll be before Oracle decides that the use of the word "Java" in JavaScript is a trademark violation and sues all of the browser makers?

[#] Sat Jan 01 2011 11:02:26 EST from dothebart @ Uncensored

Subject: Re:

[Reply] [ReplyQuoted] [Headers] [Print]

heh... probably as silly as gemaltos legal office has taken the chance to prove itself as usefully by attacking google with some patents they did about java running on embedded devices in the 90ies.

my current guess (and hopes though employed there):  those patents will go void.

and since 'java' is a word for coffee and we have trademarks here and trademarks are a little different, I think there is no big point in starting legal actions here except for over-pricing the attacked entity...

anyways, the official name is 'ECMA-Script' after all, right?



[#] Sat Jan 01 2011 11:05:24 EST from dothebart @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

oh, and now that i've decided to switch from libevent to libev for performance reasons, I've found more reasons to do so: its threadsafe and allows to have several event loops in your program in several simultaneous threads; so if libev is going to become the master of our FDs in some uncertain future, we're mounting the right horse here.



[#] Sat Jan 01 2011 20:04:11 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Spoken like a man who's never built a custom annotation with Spring
AOP.

You know me so well.


I've had a bad cold for the past few days and it's not a throat one, it's a stuffy head one so I haven't been thinking straight, but I had an idea and
I haven't flushed it out completely because I'm still sick but there's basically something wrong with code generators.
The argument starts with how fucking stupid xml is. For example why make a bandwidth intensive protocol that is human readable when most of the time a computer is going to generate it and a computer is going to read it.
Just fucking insane. Not even getting to how fucked up xml is by design anyway.

But here' my new thought:
A coworker is writing some code to call a webservice.
The 3rd party company who's service we are calling supplied her with a wdsl file.
She ran all those stupid tools and generated a fuckload of code that we'll never use so that we can call a function and wait a long fucking time for a response.

As it turned like (like every other case I've ever seen of trying to call a webservice the 'right' way (as opposed to just hard coding the soap call you need to generate)) it didn't work in production.
worked from the command line because we were able to put everything we needed (and of course there was quite a bit )on the class path.
But the webserver has it's own classpath and there's a library in our iplanet 7 webserver that supplies jaxb 2.0 and thus does not allow us to unclude a jaxb2.1 jar file so the thing just can't work.
And it's not just a matter of replacing one jar with another because the jaxb 2.0 suport is shoved in with a bunch of other things in this one big jar file and I'm not going to fuck around with the webserver setup like that.
Apparently jaxb 2.1 became standard with java 6 update 4 or something like that and we're running update 2 and of course updating is a big fucking hassle just to make this one stupid soap call.



[#] Sat Jan 01 2011 20:07:40 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

But wait I wasn't done.
Code generators.

So the damn thing never worked anyway.
But why on earth do we need a program to generate code so that we can use a nother program to compile it so it can generate xml that another computer will have to parse out.
There are so many layers of useless shit in that process (which explains why it never works) that there's really no chance of success. You have to get so many things lined up to make it work the chances of it not conflicting with something else are zero.
Everybody designs and builds these stupid tools assuming everything is always written in a sandbox isolated from all other systems where the environemtn can be defined exactly the way it needs.
And this is never the case.
Ever.


MS invented "dll hell" a decade or two ago (and I know it existed long before then) yet here we are, sun actually managed to invent a brand new technology and get it adopted by the world at large and they did absolutely fucking NOTHING to try and solve this problem.
Assholes, all of them.

[#] Sat Jan 01 2011 20:17:10 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Purely Functional Data Structures (some guy's PhD thesis, PDF)

I never understood the point of handicapping yourself with something as fundamental as storage (or rather, without storage)
I realize that I'm being hypocritical somewhat becaue you could argue back that the fact that I don't use any C++ features created after 1985 or so is like handicapping myself for no reason, but there are some obvious lines to draw.

[#] Sun Jan 02 2011 08:12:48 EST from LoanShark @ Uncensored

Subject: Re:

[Reply] [ReplyQuoted] [Headers] [Print]

anyways, the official name is 'ECMA-Script' after all, right?

You could say that ECMAScript is actually a different language than JavaScript... it's that subset which is standardized.

[#] Sun Jan 02 2011 08:17:50 EST from LoanShark @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

The argument starts with how fucking stupid xml is. For example why

make a bandwidth intensive protocol that is human readable when most of

the time a computer is going to generate it and a computer is going to

read it.

Not exactly human readable. Remember that discussion we had 3-6 months ago when I was looking at LINQ and considering using XML as a serialization protocol for query expression ASTs? You said I was drinking kool-aid, and I wound up agreeing.

Well, Siebel (now known as OBIEE) ***actually does this***, and the result is even more un-readable than you might have imagined. (They are using some JAXB-type XML binding system, so it generates xsi:type everywhere instead of using actual tag names to distinguish data types.)

Sometimes a keyword syntax (built with ANTLR, for instance) is just The Right Thing To Do.

[#] Sun Jan 02 2011 08:25:37 EST from LoanShark @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

I never understood the point of handicapping yourself with something

as fundamental as storage (or rather, without storage)

They're always slower for common algorithms... but if you get into the motivation for the optimizations this guy made, they're targeting at algorithms that choose branchpoints along the lifetime of the persistent structure and repeatedly examine different states of the structure. In other words, algorithms that need to search a solution space and perform backtracking. So if you have to analyze a large amount of data under backtracking, these might be more efficient. But for typical business OLTP problems, not so much.

[#] Sun Jan 02 2011 14:02:14 EST from IGnatius T Foobar @ Uncensored

Subject: Re:

[Reply] [ReplyQuoted] [Headers] [Print]

You could say that ECMAScript is actually a different language than
JavaScript... it's that subset which is standardized.

True. On the other hand, when most people think of "JavaScript" what they have in mind is ECMAscript plus the DOM, or in other words, the client-side browser scripting environment.

Tha language on its own is actually quite nice.

Go to page: First ... 66 67 68 69 [70] 71 72 73 74 ... Last