The sci-fi channel has been getting better and better and it’s now gotten to the point where i’m looking forward most of the week to friday, and not just for the break from a stressful job.
I got Jaimie hooked on sci-fi with Buffy. We’ve both watched every episode of that show and we’re very sad to see it end, but it was time. It had been winding down for awhile. After that I introduced her to Stargate. We’ve been watching that for the last couple of years now. Another very creative show that isn’t afraid to keep reinventing itself.
Stargate Atlantis was an interesting change. It doubles our weekly intake of Stargate and the Wraith are great, incredibly creepy bad guys. If only taila(sp?) and Dr. Weir won’t so damn annoying.
Then came Battlestar Galactica. Another great show exploring some new ideas. A little bit too stressful for Jaimie so only I watch that one.
Finally last week they introduced Firefly. Like I said I was a huge fan of Buffy and I think that a huge amount of the credit for that goes to Joss Whedon. I think he’s doing a great job with this show as well. I like that it’s taking its time to laydown the storyline. The movie comes out on Sept 30th so we’ll be spending my birthday in the theatre.
The problem with starting a business is that it takes money to do it. So for some period of time you have to live in the precarious state where money is flowing out, but not enough is flowing in to replace it. So logically you would actually be better off doing nothing at all, at least that way you’re not losing so much money. This is the scariest thing about starting a business. However, there are some trends in the software field that are reducing this risk.
In the past it could take huge sums of money to get around that corner. However, there have been some trends lately in the software engineering field to reduce that level of risk.
Sever costs have dropped dramatically. For a budget conscious startup, decent servers can be had for < $1000 running linux. These are definitely capable of getting a small business off the ground.
Software itself has dropped dramatically with open source alternatives. Linux for servers, mysql for database, java with eclipse and tomcat for applications. You can build a world class software infrastructure for $0 in licensing fees. Of course, to do this effectively you need some smart people to set things up, but if you're launching a startup you had better have them anyway.
If you're trying to reach a niche market there is no more effective way than using very targeted search ads. The cost is low enough and the rate of return high enough that as a small startup there is no need to go with massive branding campaigns to get your name out there.
There are reports surfacing of small startups getting to breakeven for less than $100k in capital. This definitely puts it in the realm of microinvestors. I believe the trend will go in this direction. The large companies will allow micro-entrepeneurs to take on the risk of developing new features and then purchase the company for a reasonable sum. This offloads the risk from the company, the entrepeneur gets a reward for taking on the risk and everyone wins.
Now if I can just get the beta release out and move on to finding customers, that's when the real fun starts.
So the first complete round of our little experiment is now over. The division was divided up into 12 groups. Each group was given a topic and told to come up with a list of quick wins for product development as well as a bit of a longer term goal.
I personally was a member of 3 groups. Issues with the process started when people began to question the scope of each group. Since the groups didn’t have central meetings and there really wasn’t anyone centrally positioned, it was difficult to answer these questions. I actually found out that two groups had each assumed that the other would cover the scope of one of the products, so the result was that no one covered it. Oh well, the joys of decentralized processes.
After a month of intesive meetings and planning each group had to give a 5min presentation on it’s quick wins. Once this was done a governing body would go through the list and sort all of the proposed items for both customer impact and engineering difficulty. The resulting grid allowed management to pick the items that were in the corner of highest impact and easiest to do.
So, a month of meetings involving the entire division and extensive planning resulted in a list of about a dozen todos for product development that was ~90% the same as lists generated prior to the process. We still do not have actual product specs to develop the products. That process needs to start now and we’ll have another month or more of meetings to get those done. I’m concerned that since the scope of these quick wins isn’t really nailed down that during the product spec phase they could dramatically increase in scope and no longer be a quick win. The one good thing is that the number of groups has been dropped down to 4 to manage the next round of the process and i’m only involved in one of those groups.
I’d have to say that so far the decentralized product development process has not lived up to the hype. It’s allowed everyone to get involved, but when it came down to it the final call had to be made by a group of managers. This resulted in people getting upset that their quick wins had been cut from the list. There was one possible idea that I know of that came up in this process that wasn’t already proposed prior. So you could point at that as a sign that we had a richer source of ideas that would had not previously tapped, but had we just remained open to including new ideas as they’re brought up couldn’t we have achieved the same result in much less time.
The final result is that the company is having a tough time handling the level of data in our day to day business. What is protrayed to the end user is our inability to develop new features for them. In each review that comes out comparing the entire field, we seem to fall further and further behind. We need to get this turned around fast or there will be serious consequences.
Last night our poor little tivo gave up the ghost. It had served us well for almost 3 years, I can’t believe it’s been that long. I believe that the hard drive was the culprit for its death. For the last day it was having trouble at certain spots in recordings. It would get stuck and then be unable to continue beyond that point. So breaking a rule that I’ve learned with servers, once a computer is up and running keep it that way, I tried to reboot tivo. It had been up for a year at this point, one of the nice things about california, no power outages. Well of course tivo never came back. I would get the powering up screen, but it would never proceed beyond that.
So, jaimie and I made the decision that we would have to replace tivo quickly, we had the tour to tape of course. Was lucky enough to be taking friday off already so that I could go grab a new one in the morning.
Tried Magnolia’s first. They weren’t really helpful. Salesguy did a horrible job. I waited around forever and could never get anyone to help me. His loss though. Went to Goodguys and the salesguy there was great. Found exactly what I was looking for immediately, a replacement model for $99 and the hard drive was twice as big now. Considered the HD version, we’ll probably be picking up a HDtv in the next year, but keep hearing rumors about those models being really flaky.
On getting the tivo home and hooked up I discovered the next problem. Tivo won’t let you record until it’s made a phone home. Of course since we’ve got voip, this was going to be a fun experiment. I found an excellent discussion of the problem here:
Since I had the RT10 model, I tried the “,#034” option and was able to succesfully make the long distance as well as the local call. Now we’ve got a beautiful brand new tivo up and running. Didn’t really realize how important it was until the thought of having to watch normal tv came back. Living without tivo would be rather painful now that I’ve gotten used to having it.
I had been trying for months, but I was finally able to do it. What you might ask was so complicated. Well, in this case it was getting my linux laptop to suspend. This had been a major annoyance that I couldn’t walk away from the laptop for awhile without worrying about the battery dying. The only way to handle this was to power the laptop down. If I closed the lid the laptop would suspend, but never come back. Seemed like the bios was trying to suspend, but it wasn’t speaking the same language as linux. I would finally have to pull the batteries just to get it to reboot.
So I was very excited several weeks ago when going through the options in compiling the new kernel when I found some options relating to software suspend. According to the brief description there was supposed to be a command to initiate this, swsusp. After some initial searching and coming up empty I finally found a post that mentioned an alternative way to initialize the process by using “echo -n “disk” > /sys/power/state”. A fellow engineer at a recent 106Miles meeting also clued me in to the fact that the mentioned swsusp command was just a shell script wrapper to this process.
So some initial testing got suspend to disk working right off the bat, but suspend to memory took some more work. I discovered that it could be a bit cranky about running processes when it attempts to suspend. Found out from the syslog that I needed to shutdown mysql before attempting the suspend. So a two line shell script later and I have a working suspend method.
Now the test was to bring the laptop back up again. With this, everything worked great except for the usb mouse. The laptop I’m working on is an old Dell C600 that I bought used off of Ebay 2 years ago. The built in mouse pad died horribly about a year ago requiring that I use an external mouse. So, after about 2 hours of more studying I found that doing a simple “rmmod uhci_hcd; modprobe uhci_hcd;” would reinitialize the usb mouse. So, now we have a working linux laptop again. Even the linksys wireless card would come back up and hop on my home’s network. Couldn’t be happier.
The only caveat with the process is that with the memory suspend it seems to get stuck in the middle of the shutdown process about 10% of the time. No real reason that I can discern and only pulling the batteries will get you out of this state. Haven’t seen this happen with the disk suspend yet, so I’ll probably be sticking with that as I only use this laptop for development work at night now. Most casual laptop usage is now on my Apple iBook which suspends perfectly, one of the nicest things about it.