Stephan Schwab

A few days back I was lamenting Is the SCM system an agile developers’ enemy?. A reader left a comment to this post saying:

…and in Subversion, you do not know who, if anybody, might also be working on the same file(s) as you, nor can you easily track changes across branches after merges. Both of these things are extremely useful to a huge class of projects, both ‘Agile’ and not, and Perforce is light years ahead of Subversion here. I agree with you that’s Perforce’s process of moving files around is annoyingly cumbersome.

So the question basically is whether it’s important to know who is editing a file before you can open it. I’ve used other SCMs before and I remember doing phone calls to other people in order to ask them to check in a specific file so that I can open it for editing.

My point is that software development is something different than working on configuration files or documents. In many cases I agree that it’s a useful feature to let only one person at a time make changes. But to a software developer that’s not entirely a good thing. Where is the problem that Joe is making chances to method foo() while Jim is doing something to method bar() that are both in the same file? That happens millions of times a day in every project and is completely normal. What’s needed to deal with this is not a tool that limits changes, but a tool that makes merging easy.

Joe and Jim should be allowed to make changes all over the place with no restriction at all. That way they can work away and experiment with different designs, which allows them to find the best solution for the problem they are working on. When each of them is done and wants to share the new stuff with the rest of the team, they tell the SCM to compare their changes to the version of the files in the repository. The SCM should show a side by side view of their version of the file and the one stored in the repository. Now it’s easy to decide what to move or where conflicts are. If there are conflicts those can be dealt with before the next merge attempt. Have a look at one of the major Java IDEs (IDEA, Eclipse, Netbeans) and you find exactly that kind of merge editor I just described.

Again, there is no need to prevent people from editing the same file in a software development team. Software development is a creative process and whatever limits or restricts creates a boundary for the creativity. The result of imposing limitations by using the wrong tools will be software that looks and feels just like the limitations its creators suffered from. Please read about The Windows Shutdown crapfest to see how such a thing can look like. It should open eyes.

From my last trip to the US I brought home a brand new Apple TV and connected it to our big screen. It instantly found my WiFi network and started sync’ing to the Mac Mini that was connected to the TV before and now has moved to the home office. The Apple TV fulfills its duties as a big video iPod quite well and my wife loves it to not only watch movies and TV shows we download from iTunes, but also to fill our living room with Salsa rhythms. Sure it’s a bit unusual to switch on the big plasma TV when all you want is listen to music. But on the other hand the cover animation used as a screensaver is nice and makes us switch to another track because we remember it by seeing the cover of the album it is on. So all in all the Apple TV is a nice addition and it just works.

Now reading my feeds I came across this:

Alasdair Allan in The Daily ACK: The new iMac:
Want to make a bet that that that refresh for the Apple TV that Steve hinted about during the Q&A yesterday means that you can use this new lap-friendly lightweight keyboard to do a lot more than you used to be able to do with the Apple TV than you can now?

Here is an image comparing the two keyboards:

The prospect of using a wireless keyboard with Apple TV is compelling. Maybe that “video iPod” gets extended to include at the very least Safari to browser web pages on the big screen. Then such a small keyboard and a wireless both would make great sense. Unfortunately the technical specifications for the current Apple TV doesn’t list Bluetooth and I doubt a bit that it’s already integrated, but deactivated. We’ll see ;-)

It is interesting to see that more and more people are thinking about serving the needs of Spanish speaking people within the U.S. and in Latin America.

François Schiettecatte in GigaOM Introducing, Boston Start-Ups with Hispanito «:
Derely says that the Spanish speaking community is very under served and most of the social networking sites are targeted to English speakers. The popularity of Fotolog, Google’s Orkut and Hi5 in some parts of Latin America indicate that targeted consumer services focused on the Latin American market can become thriving businesses, even though only 20% of Latin America has access to the internet. The internet connections are growing at a rate of 30% a year. At present there is about 8.2 million broadband connections in the Latin America. Derely feels there is a lot of room to grow but it will take 2-3 years to really pan out.

As readers of this blog probably have noticed I live in Panama and run a software development business there. Panama and other Latin American countries have a lot of talent interested in this line of work. And of course there are a lot of potential customers. The only problem I see at the present time is that on one hand people don’t have the money to buy a lot of services and when they do they can’t pay by credit card. So there is a need for other payment mechanisms and services have to be cheap. Further one should keep in mind that the majority access the Internet from public computers in Internet Cafe’s for about 50 cents per 30 minutes.

It will take quite some time to come to the point that a significant number of people – first of all – own a computer and then can afford broadband Internet. For example here in Panama a 256 Kbit/s (yes, it’s that slow, but still called broadband) connection costs about USD 25 monthly. For the same price you can have dinner for two at a nice restaurant. To put that even more into perspective: most people earn a monthly salary around USD 600. It really is not that easy to afford a computer with Internet access when you need the money to make ends meet.

The opinion, as expressed in the cited article, that it make take 2-3 years to make a service grow successfully, which means generate enough revenue, might be a bit too optimistic. The amount of disposable money won’t come up that fast in Latin America.