A Long Cold December

topcoder: confused??

Sat, 28 Aug 2004 21:45:31 GMT

So in the middle question I got confused and completely screwed up.

I went back this afternoon and took a look at the "hard" question... I finished it in something like 20 minutes.

no problem.. I understood exactly what I had to do.

wtf?

I think I may be taking this from the wrong end

I've been going in starting with the easy questions and working my way up. Maybe I should start with the hard questions and work my way down.

sounds bloody stupid, but I usually score better on the harder ones then the easier ones.

going to india

Sat, 28 Aug 2004 17:53:57 GMT

Okay, so we've been outsourcing work for years and years.. Canada, Ireland, it's all the same.

Recently we've been making a push to outsource to India just to see how that would work.

I tell you theres as many cons as there are pros to the situation.

I'm overseeing the rehosting of one of my projects to the weblogic platform, as part of that rehosting we are doing a complete redesign. ( fun stuff too.. You should have seen the architects face when I mentioned he left off some zero's on the file transfer capability ... "no it's 30,000 files a day with a max file size of 80 gigs" );

Part of that oversite is to perform the code reviews. That poses a problem, we can get one or two of the developers from their team to fly over here for the review, then hold it, and possibly miss out because it's impossible for a respresentaion of a development team to understand all of the code.

Or, the obvious better choice ;), would be for my company to fly me to india and hold the code reviews out there.

I am pushing for the trip out there.

Since it's so far my company would automatically book me first class as well. :)

Life would be good.

topcoder anxiety

Sat, 28 Aug 2004 17:47:25 GMT

topcoder is really bothering me.

If I was to base my skill set on my topcoder ranking I would apparently be a dumbass.

I know without a doubt that I am a good programmer. I am consulted on a regular basis for my opinion on not only code but design and implementation. I have dealt with architects, senior developers, a couple of people who have gotten their shiny little faces on books about java development and I can stand toe to toe.

Yet I consistently abuse myself by doing the topcoder competititon and doing miserable.

It shakes my ego.

I know, in part, its because it is aimed at a certain type of developer. Those that have studied algorithms or do it as part of their job.

I've never had to write a complicated algorithm in my life.
Hell, I've never even taken a computer science course.

Sucks to be an accidental developer I guess.

I am getting better, I've figured out that theres a certain style to the programming that helps.. I've had to change my mindset to forget such things as design and re-usability. The first time I did the competition I was so stuck on figuring out how to design the solution I ended up screwing myself over.

It's just an "arrghh" situation.

bahhh.. anyways on to better and moire entertaining news.

eclipse

Sat, 28 Aug 2004 17:33:16 GMT

So I spent the last week or so evaluating eclipse as a base development package to build gui's on.

It's allright I suppose. I like the way they implemented what the call jfaces over the some of the widgets and their gui really does have a nice feel to it.

I've worked in java's awt on and off for the past year or so. I've done custom gui's and one or two games and even a ripoff of winamp (really should finish that one)

You can develop really nice responisive gui's using java. but the biggest problem is the amount of work it takes. Swing, IMHO, is just a real bear to work with and is often way over complicated for what you need.

In that context the choice of using eclipse's swt over awt is a clear one.

I guess the ambivalence I feel is with the way they implemented what they call the rich client platform. I'm not going to tell one of the devlopers not to work in it if they want want to. But theres no clear association over how some things work.

I think thats what bothers me about a lot of containers.

If I write a piece of code I can see clearly the interaction between one piece and another. a causes b causes c. With containers, I can build a piece of code over here, another piece of code over there and somehow they interact, sometimes because of defined interfaces and sometimes because you've stated that they will through an xml document.

I feel ill at ease with that type of development. I know, I can lay that unease on my own doorstep, but still I can't help being bothered by it.

But yeah I'm going to reccomend eclipse as the default basis of any management gui's we build.

We'll see how that works out.

something else I want

Wed, 11 Aug 2004 10:35:26 GMT

A small application container.

j2ee sucks.

I think j2ee will be next to xml in the future history books of dumb ideas that everyone and their brother bought into.

It does have some useful properties (which is one of the reason people buy into it) the whole container concept and providing nifty things like database pooling. Thats useful. But for the vast majority of projects I work on j2ee is overkill and yet I am forced to deal with one level of abstraction of another of j2ee on a daily basis.

What I need is a small container that doesn't do scaling, that every object in it is local to the other objects, and that provides an easy to use configuration interface.

That would make my day.

back again

Wed, 11 Aug 2004 10:23:11 GMT

So my daughter went home to her mother on Sunday.

Gives me a chance to focus again on some of the fun stuff, or at least take my mind off of things.

So in my spare time I'm working on making my own programming language. It's one of those things that you do for the hell of it just to see if you can do it and to also learn a little bit on why things were done they way that they were done.

The major decision that I made was that this would compile to java byte code, since thats the language I'm most familiar with. It also gives me a slew of stuff thats just plain useful, portable code, a pre-existing library.. etc.

I almost bogged down on the project when I started into looking at what I had to do to build the class files. At this point the majority of options were overkill.. I wanted to focus on the design and have a result.. not spend my days building a compiler.

So I stole a trick from PERL and a couple of other applications.

I'm going to use ANTLR to parse and tokenize the program, then actually have a .java temp file created write the equivalent of what my language would be if you wrote it in java and then just use the javac to compile it to byte code.

no mess, no fuss, everyones happy.

A time for a change

Tue, 15 Jun 2004 14:47:46 GMT

I came to the the realization recently that my job is stiffling me.

I became a programmer because I enjoyed problem solving and creating things. That part hasn't changed. But I've found myself in a departmetn where very little orginality is needed. I think thats starting to sink in.

I need to do more stuff where my ability to come up with orginal solutions is required instead of the odd occurence where it helps out.

Right now. I do internal infrastructure and support applications for my company. I need to find a place that is actually designing new stuff. New applications, new features, that sort of thing.

around here we call that a royal ass kicking

Sun, 13 Jun 2004 01:15:09 GMT

Oh don't get me wrong, the things that happend earlier today was annoying but after going back in and evaluating what I had done and then looking at some of the code that was written and I realized at this point I got out classed and body slammed.

*grin* the word is humbling.

It reminds me a lot of word puzzles when I was a kid.

I'm thinking a lot of it is has to do with having a clear understanding of whats being asked.

visualization so to speak.

Time to start doing some home studying.

topcoder contest #1

Sat, 12 Jun 2004 20:26:17 GMT

You ever have something go so wrong that you can't wait for the next time to redeem yourself?

Issue #1 Was that parents decided to pay me a visit 3 minutes before the start of the contest. I know, I know, it sounds like a bad sit com or something but I'm like "just sit there for half an hour and don't say anything." So being parents they are wandering around cleaning the house, looking over my shoulder to see what their son is up to.

Issue #2 Why yes their is a difference between just going online and looking at the questions and actually being in a round. Nice little adrenaline rush going on. Especially cause I was screwing something up.

I took the easy question because I thought I should get use to the whole system first.

That would be the reppity :) I kept getting sequence results rather then substring results and I couldn't figure out why *ha* God that was frustrating.

So I was pretty sure I fixed the problem, ugly as the code maybe, all the tests came out right. Submitted it and then took off to take care of my unexpected guests. Came back and discovered I had been challenged and all my points had disappreared. I'm not quite sure how that works and I have no idea how the challengs occurs but I can't help wondering what I did wrong *chagrin*

So I'm grabbing my code and reworking it and stuff and I'm going to spend the rest of the week working through the practice rooms.

It's nice to have a challenge again.

topcoder

Fri, 11 Jun 2004 19:21:59 GMT

I just registered today at topcoder.com. Topcoder is an online contest where you can compete against other programmers for cash and prizes (I sounded like a blurb there.. scary.)

It looks entertaining and I'd like the chance to see if I'm as good as my oft inflated ego thinks I am. The only drawback of course is that I will, in some way, have to deal with other people. Having social anxiety really sucks ass at times.

The tasks that they give the competitors to perform don't look that particularly difficult. However I'm not sitting there in an artificial arena watching the clock tick by either.

Adrenaline is good for the soul.

There's a competition tomorrow. Since I don't know what to expect I'm not going to be to hard on myself. But I am going to spend tonight and tomorrow morning going over old tasks and see how I perform, that should be entertaining.

sax driver impl xmlrpc

Fri, 11 Jun 2004 19:08:09 GMT

Playing around with sax for my xmlrpc implementation.

Right now I'm writing it with the emphasis on the trailing edge (is there a proper name for that?). So when I receive an event for a start element all I'm doing is hashing the name and placing it in an int stack. When I get an end element I first hash it and check against the stack to insure that it's well formed xml. Then, based on the type of element, I perform some sort of operation.

It seems to be working very handily. I'm quite pleased with the times that I am seeing. Now I need to build junit tests against it and compare it against other impls.

I'm torn sometimes,making things go fast is fun but I have this really hard core desire to write by design. I like to see code thats minimalistic and beautiful. Call it a failing.

Confined by our forefathers

Wed, 09 Jun 2004 21:59:16 GMT

As programmers we face an artificial constraint that is placed on us daily.

That constraint can go by multiple names: ascii, english, or the keyboard. Either is as good as the other, they are just reflections of our embryonic
programming practices.

Look at other fields, as you advance through the complexity that they invoke you will run across symbols, theta's, pi's, functions, all of these were required because ordinairy words had begun to fail. There is an inherent limit on what can be shaped by the simple letters and numbers of our language, so special symbols were invented and/or borrowed to allow us the flexibility to represent large and complex ideas in a simple format.

Except for the field of programming, in programming we are stuck to the letters on the keyboard. A keyboard that was developed for something very different in mind then programming. Letters, basic communication, books.. yes a keyboard is fully functional 99% of the time for these endeavors... but as the basis of a programming language?

Hah.

So to compesent for the weakness we take the existing characters that are on the keyboard and make them do and represent ideas they weren't originally made for.

That initself is not a bad idea. However we aren't adopting symbols because they are the best fit. We are adopting symbols cause they are all we have.

don't see a problem in that yet?

I see patterns. It's what I'm good at, in fact I excel at finding patterns and I can guarantee that one day the killer programming language will be one that is composed of a pallete and a variety of interactive symbols that will allow us to build what we need in a manner that can not be represented intelligibly by text.

The pallette will interact with our design, it will be intelligent, it will support functions and methods, and we will quite literally be able to paint programs rather then be forced to rely on the keyboard, We will be able to design loops and functions and understand their operations at a glance because symbolism is power.

I'm one impatient person.

BEA Weblogic and BEA Workshop

Wed, 09 Jun 2004 16:58:36 GMT

it's because I open my mouth and mention that I can do it.

I know thats the reason.

I'm not to happy with workshop. For the majority of things that I have to do it works just fine, and is, a fine fine tool. However, stray outside of what it has established and you are stuck in a quagmire of pain.

Thats the only way for me to describe it.

Their api's are so convulated and screwed up. I spend half my time working around them.

Still I end up doing things faster in workshop then if I was off doing this solo. So it's something that I have had to suck up and accept. Not particularly sure if I'm happy with that fact.

xmlpull

Wed, 09 Jun 2004 13:34:43 GMT

I really hate xml. Yet still I work with it. I wonder if it's some sort of inner masochism I have going on.

The xmlpull api (http://www.xmlpull.org ) looks rather sweet. It's a streaming tokenizer for xml. Which is the type of setup that I have some prior experience with (that would be the time period where I went mental and started building my own parsers and languages)

Tonight, when I get home, I'm going to implement it into the driver framework for my xmlrpc lib.

*certain level of joy*

xmlrpc

Wed, 09 Jun 2004 12:41:19 GMT

I'm currently working on my own xml-rpc library.

Yeah, I know there are libraries out there, but two things.

1) they aren't mine

2) I like to experiment.

So I'm using this opportunity to explore different ways of implementing parsers etc.

Actually, doing some reading yesterday I'm going to spend the day looking at xmlpull to see if that would be appropriate for the type of messages that I deal with.

a note to oneself

Tue, 08 Jun 2004 13:37:48 GMT

I was browing the algorithms forum when I chanced upon an entry where a person was having issues with a test question. Make a program that prints out the numbers 1 to 100, then 100 to 1 without using an loops (ie for, while, do) Most of the responses were interesting but all were in c or a variant and used a conditional "if" to determine when 100 had been reached.

As a matter of self entertainment I wrote a java method that performed the required work and used no control flows.

public static int recurse(int value){
System.out.println(++value);
System.out.println((value<100)?recurse(value):value);
return --value;
}

so there.