Banal Minutiae

Another fun weekend gone.

I’d have to say the best thing about weekends is the complete lack of having to do anything. I personally consider work to be anything I don’t want to do, so the weekend is a prime opportunity to do things I do like, and hence, no work.

Even working on my house can sometimes be fun, but that’s another column.

My DVD player is amazing. It has an ability I had never seen before – the ability to play MPEGs and AVIs. Amazing. Basically, create a DVD full of these files, and it will play them. It has an explorer-like interface that lets you navigate the files. It’ll also do MP3s, OGG, and WMVs. Basically, anything you can imagine.

Why is this awesome? Well, if you’ve ever downloaded a missed TV show from the internet, you know what a PITA it can be to watch the show. You have to encode it in the right format, make sure the audio track still syncs up, etc… Additionally, it takes forever, which in my case is about 4 hours of computer time to encode an hour and a half movie. Basically, way to much time. Basically, even after all this, you still need to burn it to a disk.

With this DVD player, all you need to do is dump all the files to a disk, and bang, there you go. Not that I’m advocating this or anything, but if you have a show you really like, say 24, you can dump the entire first season onto two disks. Awesome.

This works great with my Tivo, which is hooked into my network. Tivo2go allows you to transfer files to a PC, which come over in a lame copy protected format. I’m sure if you have at least a brown belt in google-fu, you can figure out how to convert a .tivo file to a straight AVI. This is great, as my Tivo has limited storage. It’s also nice to go back and watch a complete season of a show. If I’m really into it, I’ll even edit out the commercials. Works great.

Some friends of mine and I are into the DVD swapping thing. One guy loves the old cartoons, and recently let me borrow Star Blazers. Man, that takes me back. It’s basically unwatchably horrible, but can you really put a cost on memories? Lost on the other hand? Very nice. The only problem with it is that similar to 24, you don’t want to miss anything. It’s painful to just say “One more episode”, and realise you’ve blown like 7 hours sitting in front of the idiot box.

Anyway, work sucks, but you already knew that.

Since I seem to be ranting about work, let’s talk about VSS, Microsoft Visual SourceSafe.

Where do I start?

In case you’ve never heard of it, it’s a piece of software that does software configuration management (CM). In case you’ve never heard of that, well, you’re not a programmer. Or at least one who’s worked on a big project.

In a nutshell, CM basically gives you control of a big project, and allows multiple people to work on things simultaneously, while allowing for the all important tracking of changes. It’s a program, much like Windows Explorer, where you log into a project, and see the entire project hierarchy. You can see which version each module is up to, who checked it out last, and if any modules are currently checked out, as well as who’s working on them.

This all stems from the common problem of multiple people wanting to edit the same file at the same time. It just doesn’t work. After all, you can’t have one person undoing or missing changes another just put in. This brings us to the complex process of branching and merging.

Branching and merging is another common process you need to understand. Basically, with branching, you start with a single file or project, but then create two or more projects from the original. You leave the original project or file alone. Work proceeds on the branched projects until they are complete. Now, you have at least two different files or projects based on the original. At this point, you can merge them together to create a new file or project with the attributes of both branched projects. This can be tricky, as there can be conflicts in the code. The advantage of this process is that it allows you to develop major new features into your project independent of one another. This means that before you integrate all the new features into your code, you can already have made sure each individual change is working. Then, with merging, you can resolve feature conflicts. Got it?

Anyway, my original rant is on VSS. You see, VSS is just not a good system for massive parallel development. Why? There are so many reasons it actually becomes hard to know where to start.

Let’s start with the bug most likely to bite you on the ass. When you check a file out of VSS, you copy it out of the database to a physical location. In order to compile, we copy to a network location (it’s really not feasible to work off your own machine). Now, here’s the thing. If someone else edits or overwrites that file on the server, you’re screwed out of all your changes. I can’t tell you how many times this has happened to me. I always keep local backups of my changes, but you still have to work off the server for space reasons.

Another problem with VSS has to do with the differencing program. When you do differences between files, it uses your last defined temporary directory. Again, the problem here is if someone else has used that temporary directory. If they’ve modified your file in any way, you’ll get false differencing results. Crazy.

Look, I realize that these problems relate to our unique setup, but another problem has to do with the limitations on database size. Our code base is rather large, and we’re just about hitting the limit. Another problem is that the database is in a proprietary format, and if it gets corrupted, you’re basically hosed. Both problems have hit us before.

We really need to move to a large enterprise sized CM software solution, but management doesn’t want to invest the cash. Typical. I really don’t understand this argument. You see, we all spend extra time grinding our teeth and dealing with scalability problems stemming from the limitations of VSS. If we upgrade, we pay a purchase cost, but then the lost labor would go away. I guess that doesn’t matter.

An old expression (and somewhat offensive), but I think quite apropos to my day here.

I mentioned before that I’d been working on a ROM. Well, I finished it this morning, and was looking forward to going back to working on my ATP that I did earlier in the week, when I got dumped on to look into a potential problem.

It’s crazy I know, but the general attitude around here when a potential problem is found is to throw as much labor as possible at it. This is good and bad. It’s good in that lots of people can work on differing aspects of the suspected problem, but it’s bad in that everything everyone was working on is now sitting idle, and with the stuff I do, there’s definitely a ramp up time associated with going back to work. In other words, you usually need an hour or two to get back up to speed. That might not sound like a lot, but I assure you, if you get hit with four or five interruptions, it starts to add up.

I’m not really complaining, because the potential problems need to be addressed. I’m just saying it’s a problem that bugs me. I guess I am complaining. It’s about something that can’t be helped, so it’s not a big deal, it just feels good to talk about it.

Of course, the real, underlying problem is the code base we work from. It was started in 1982, and hasn’t really seen any updates. It was ported to NT 4.0 in the late 90s, but the code itself was just ported to C (from PL/M). No improvements were made.

Basically our code is seriously outdated. The major architecture is based on a modified semaphore or mailbox system. Yikes. Well, back in the day, there were memory concerns. A program couldn’t be larger than 64K and still fit into memory, so the mailbox system was developed allowing smaller modules to exist, operate, and communicate. It’s a problem, because it makes things really complicated.

There’s also the fact that the older code doesn’t even have simple things like matrices. That’s right matrices. Back in the day, only one dimensional arrays were around, so a function was created to mimic two dimensional arrays using a one dimensional array. It’s really inefficient.

Speaking of inefficient, would you believe there’s no pass by reference? It’s all pass by value. Wowsers.

I guess I’m on a bit of a rant. Look, as far as I’m concerned, we need to think hard about rewriting our core architecture from scratch. When you look at the bottom line, I think there are serious questions about development time using our current code base, and the problems that keep popping up. Does it make sense to continue on our path, knowing that even simple tasks are going to take more and more time, or does it make sense to take the hit once, and then have smooth sailing, as well as vastly improved confidence in the code base?

Real world answer? I don’t know. It’s obvious where I want to go, but I’m not the one in charge of the money. That’s reality. The people best able to answer questions like this are seldom asked.

Man, I’m bitter.

Well, I’ve been working on a ROM (that’s Rough Order of Magnitude for those keeping track at home) proposal for some work for Bet Shemesh engines over in Israel. I’m not a big fan of proposals, mainly due to the type of work we do.

See, we’re in the research and development business. That means that people ask you to do X, and sometimes you don’t know how you’re going to get there. Sometimes things come up along the way you couldn’t have known. All this means that you might guess something will take a week to do, but it actually takes a month. Management doesn’t like this scenario. It could also go the other way; something you thought was going to be hard might be easy. This is bad because your customer rate is going to be artificially high. Bad.

In any case, one of the problems in dealing with the Israelis is the language barrier. Lots of times there is confusion over what exactly is meant. Another problem is just that the Israelis seem have a completely different business ethic. Everything should be free with them. Contrast that with the US government, which pays for everything. Hard to accept the former when you’re used to the latter.

Anyway, my point. I hate dealing with this stuff, because it’s really your butt on the line. Unlike some other people in my group, who bid projects that are “a lock”, and can thusly pad their hours, this is really more of a competitive bid, and it’s not at all sure we’ll be getting it. Because of this, I need to be leaner with my hours. It’s a fine line. Go too lean, and you risk disaster if there are any problems. Go too fat, and the customer will never go for the work.

Some people have this down to an art, but I’ve been doing it a long time, and it seems the harder I look at this stuff, the more roadblocks I see. I’ve definitely gone to a pessimistic attitude over the years on this stuff. I’m sure that’s in some way related on getting burned on a couple of my early proposals, but that’s a different story.

More later.

So, yesterday after work I went and checked out what my options were in cell phone land.

As you can imagine, this is just about as confusing as buying a home. I had to stop into Best Buy to return a not so universal remote (another story for a later time), and I noticed they have a cell phone area that includes Cingular (along with everyone else.)

My limited research had led me to believe that the middle of the road Motorola V551 would be a decent choice. Best Buy was selling the Motorola V557 (apparently the newer model) for $9.99. Damn. New front runner. I should have signed up on the spot.

Instead, I decided to go over to the Cingular store and see what kind of deals they had. Turned out, a more accurate assessment might have been what deals they didn’t have. They had the V557, but for $150, with a two year contract. The guy there tried to convince me this was a good deal, and their better service more than offset the price difference I’d get at Best Buy. Riiiiiiiight.

Option two would be to switch to T-Mobile. I could probably keep my number, and since they’re GSM on the more international 900/1900 bands, I’d have more phone options.

I think I’m leaning to sticking with the $10 deal at Best Buy. Cingular has been good to me over the years, and I don’t have any major complaints. Coverage is also good in all the areas I travel to, although I really haven’t been doing a lot of traveling lately.

I’ll post later today on what I end up doing.

At this point, I hate my cell phones. Yes, I have two.

I have a Samsung S307, which I’ve had for almost four years now. Here’s the problem with it. The speaker isn’t loud enough, and the microphone sucks, so people I talk to are always complaining about my call quality.

About a year ago, I bought a Sony Ericsson P910. It’s a combination PDA/Phone. I got it thinking it would combine all the cool features of my Pocket PC with my phone. Here’s the thing. It’s quite tantalizingly almost great. And by that I mean it’s the most annoying device I’ve used in a long time. Why?

Here’s the good. As a phone, it has a great speaker and microphone. Both rock the Casbah. It also has bluetooth, a cool voice dialing feature I never use, a camera I use occasionally for those humor snaps in my day to day life, and a speaker phone I use a lot. It has a somewhat clunky dialing interface. PDA stuff is pretty good. I have a MAME emulator that plays lots of cool games. It’s pretty sweet in an airport or a waiting room. Web browser is amazing. Email – same deal. Cool. It’s got a keyboard, so you can really bang out useful stuff on here.

Here’s the bad. It locks up. That’s right, my phone locks up. Freakishly annoying. I can’t tell you how many times I’ve pulled it out of pocket to discover it’s frozen, and I missed a call. Also, the calendar app likes to sometimes forget to remind me of appointments. I’ve considered smashing it for that. That’s the main reason I bought the damn thing. It also has sync issues with Outlook. Lots of times, a contact will duplicate about twenty times.

These are all really minor compared to the number one deficiency: the battery life. My Samsung will go for days on a charge, but I’m lucky to get a full day on my P910. Maddening. A half hour call will burn 50% of your battery life. Ouch.

One other minor issue is that the P910 is big. The Samsung is too small. It’s hard to hold.

So, here I am. I’ve been carrying the Samsung right now (I can just swap my SIM card), due to the battery life, and the fact it’s easier to carry, but I need a new phone. I’m sick of the pathetic volume and call quality on the S307.

I’m considering a couple of phones. I could go with a standard Motorola V540, which is just your everyday clamshell. Nothing fancy, but good in every respect. I could get a Motorola A630. It’s the little one they carried on NCIS last season. Got a flip open keyboard, some PDA functionality. Both are pretty cheap on eBay. I don’t know right now.

I’ll think about it some more.

Ah, Friday, glorious Friday. I love me some weekend. It doesn’t get a whole lot better than no work.

No work is the adult equivalent of no school, with the benefit of no homework. Sweet. Of course, in a stunning turn of events, there’s the dreaded mortgage, along with various other sundry bills. Hmm. Homework is looking up.

Just sitting here doing some data collection and analysis, and was thinking how much the weekend is going to rock. I don’t even have anything planned, I just know that being here is better than being here. And that’s the bottom line.

Hey! First Post over here on WordPress.

I’m contemplating moving from Blogger to WordPress. Just seeing what the other side of the fence is like, as it were.