Fog Creek Software
Discussion Board




Lord Palmerston on Programming

Joel Says:

"So for now, my advice is this: don't start a new project without at least one architect with several years of solid experience in the language, classes, APIs, and platforms you're building on."

This really does not work with new technologies like .NET, even though it is good advice, I think that it is more applicable to C++, Java, VB.  Hopefully your are able to find a decent architect with years of doing successfull architecture in C++, Java, etc.  Given what they've experienced, they should have good insight into what to do and what not to do.

Tim
Tuesday, December 17, 2002

I'm following the policy of not using .NET on anything critical (certainly not launching new products built on .NET) until we've collectively got enough experience with it from small, non critical, in-house projects.

Joel Spolsky
Tuesday, December 17, 2002

Actually, the problem of developers having a good year of .net development is a problem.

Right now, the industry as whole is sinking its teeth into .net. It is a real big dog bone to chew on.

As a result, the industry as a whole is still about 1 year always from really seeing the effects of the .net environments.

Once the developers really get up to speed (and learn what works well, and what does not), then we will start to see much more fruit form .net platforms.

When it does happen, Microsoft will have won a incredible war here. Imagine the guts it takes to change a whole platform. Microsoft could have just done nothing. Microsoft is one big bold company. No wonder they are where they are in the marketplace today.

However, in the meantime, there is some struggle as the industry tries to absorb this big pill called .net to swallow.

There is also the issue of the existing software base..and it grew by huge amounts in the last 5 years (and those were not .net products).

While the industry struggles with this change. I think it will be the last time that the industry accepts such a large change from a single company. Next time around I don’t believe that Microsoft will be able to accomplish such a incredible feat.

As a society, we cannot for example decide to throw out cars overnight, and replace them with something else. The same concept applies to the computer industry: We cannot afford to change everything simply because a new technology is better then the old. The existing software base is very large right now (larger then it has ever been).

As technology matures, it gets better. It also becomes more difficult and specialized. Hence, the investment it represents also increases over time. As the technology gets better, it also lasts much longer. At the end, it means the stakes get higher and higher if a change is to be implemented.

40 years ago, a new fighter jet came out every 15 months. Today, a new fighter jet will probably LAST 40 years!!!

The software industry is the same.

Hence, the real question is the cost of change, and the maturity level.

As the software base gets larger and larger (and matures), then the cost of changing also becomes more and more difficult.

While we brought structured programming in the 70’s, and concepts of objects came in the 90’s we did not make such a large change in the tools we use.

This current change we are making is going to be a real long date and romance with .net.

Albert D. Kallal
Edmonton, Alberta Canada
Kallal@msn.com

Albert D. Kallal
Tuesday, December 17, 2002

The programming platform is only one facet of a development project. It is an important one, but is not by any means a silver bullet for development woes. So to prematurely give .NET credit for revolutionizing an industry riddled with non-platform-based problems (as well as platform-based problems) is ignoring the other hard stuff.

Regarding whether or not software gets better as it matures, and thus lasts longer I'd like to question whether that applies to Fortran, Cobol, etc. Systems built on these old technologies still work, but I don't think anyone would necessarily make the argument that Cobol has lasted so long because it has matured. It has lasted so long because it still works and because of the investment made in it.

Finally, I agree that MS is big and bold, but I would put neither of these terms in a positive light. They did not "embrace" cross-platform functionality until faced with a competitor who seemed to be making inroads based upon it. They rarely innovate, but instead follow and oust the trailblazers through questionable (if not illegal) business practices.

Jeff Kotula
Tuesday, December 17, 2002

*  Recent Topics

*  Fog Creek Home