Reuters reports that IBM will be building a 1,186 node, 2,300 AMD 64-Bit processor cluster supercomputer for the U.S. Army. The computer is expected to clock-in at 10 teraflops, and is expected to be in the top 20 of the top 500 list of supercomputers. A recent article in Slashdot spoke on the alleged crisis of the reduced use of "real" (and much more expensive) supercomputers vs. "multiprocessor matrix" (my term) clusters such as these.

A news item regarding Open Source Myths appeared here on Slashdot about the article on Neil Gunton's website to which I wrote a reply. I made some comments to him in response. I have amplified some of those comments and expanded upon them here.

Title: "I liked your comments about open source"

And I happen to agree with the points that you have made. But I don't necessarily believe that things are as bleak for the small developer as you make them. You have just as much risk that if you develop something successful that some larger company may develop something similar as well as someone else developing an open-source product that might also compete. Certainly it be nice to be someone who wants to make money writing programs and not "spend all his time on the phone doing support." But unless you are working for some large company that can hire the people who they need to do that stuff, you're going to have to do some support yourself of your product. I do it with the stuff I write.

But you take that risk when you enter any business. The buggy-whip manufacturers and horse carriage makers were in a great line of work until the automobile killed them. Some of those companies switched over to building parts for automobiles. They adapted to the environment.

I also think you're throwing the blame for the inability of a single programmer or small group to develop some really hot application and sell a huge number of them to a lot of people is unworkable today because those who got there first have raised the bar by increasing the expectations of users to include a large number of additional features and "bells and whistles," many probably having little to do with core functionality, but still raising the cost of development and increasing the complexity and cost to support.

While some of this is a good idea, a word processor that provides all of the functionality to write a complete book is something I needed when I started writing the books I'm working on, there are lots of things in Word Perfect that are of little or no use to me, but may be of use to others. (I'm not one of those Word Processor bigots, I've tried Microsoft Word, and in discovering that it could not do what I wanted for my books, in examining one of the mistakes it made in formatting I decided I liked what it did better than what I was using, and was able to quickly borrow the feature back into Word Perfect.)

In the response to you I had forgotten to add - and I'll do it here in my journal entry - that it's estimated that only 5% of all software developed is of the "boxed retail" or "intended for public consumption" type application. Most is for in-house use, is developed for private release or for a specific customer, or is developed for specialized applications such as device drivers and software to run appliances. In none of any of these cases is there even a commercial market for the software being developed, and there is no "external price" available for a large percentage of software being developed because it is stuff that is developed for non-public consumption.

You seem to knock going into "niche" products or into things that are not very popular, or going into custom development. Some of these things can be quite lucrative, especially if you get into woefully underserved areas where there is a big lack of a good product for a particular marketplace. Because if it's a small market that doesn't have a good application to solve its problem, there may be room to charge more for solutions that are not being served.

I think what you're complaining about is that the "low hanging fruit" and the "easy" stuff that was very lucrative without a lot of effort has already been taken. Whether that's true or not, I don't know. But I do know this: trying to get by in a market by going after the "easy" and "simple" stuff is a sure way to be marginalized by others who have the same idea (or, as in your examples, to be usurped by some open source application.)

But let me push your own words against you: If it's not something that will "scratch the itch" of some open-source developer, you're not going to see an open-source application come out in competition to it. Some kid in a garage is unlikely to develop an open source payroll application to compete against some commercial one, or we should have seen open-source payroll packages by now. It should not be that hard to do, there are even plenty of source code modules in existence going back to the 1970s when the DECUS source code libraries were released for PDP-11 and other Digital Equipment Corporation machines. So it's not like someone even has to do this from scratch, a lot of the work has already been done.

I have yet to see open-source banking software. I don't mean programs to run one's checkbook, I mean ones to run a bank. Or a credit union. There are lots of small credit unions that probably do a lot of work with paper-based files and manual transaction work that could probably run their operation on some sort of networked package which is not expensive (compared with the huge prices I'm sure large vendors charge in comparison.)

I'm guessing here and may be completely wrong when it comes to the above example, but I'm sure there are lots of other industries and applications that could use better software tools to support their business and could pay the kind of money that is worth it for a single-party developer or small group to get involved in, but not enough for large companies to bother with, and too much trouble for some unpaid programmer to get involved in developing an open source competing product.

Now, looking at, for example, financial applications, while there is a development of GNUCash, a clone for Quicken (which basically is a system for managing personal finance, something an individual as a programmer might be interested in), I have yet to see any interest in developing complicated financial software for businesses such as the complete package (payroll, Accounts Payable, General Ledger, Trial Balance, Accounts Receivable) even though it's been done - with source code publicly available - in other areas, as I noted above.

That's just one example.

Simply, there are lots of different application types individuals and small groups can work on for sale that are of little or no interest to the "scratch your own itch" type of programmer who might be tempted to write an open source application to compete with something you might write, if it's of the type of material that they have no need for.

Just think of other applications that programmers have no need for, and you're unlikely to find someone else coming along and writing an application in competition to it at the free-software level.

Also, most of these people involved in stuff in that environment tend to migrate to Linux; very little of it filters back to Windows. So if you hang around Windows applications it's unlikely to filter over.

Another thing: push for and strive for better interfaces, easier to use and more intuitive development for the non-programmers who might use such applications, and this will also raise the bar to the open-source movement in developing competing alternatives to what you might create. They are extremely weak in developing good quality software that is easy for the average person to use. Their usability factors are often very weak. If you target the parts they are weak on and market to those, you can often stay in areas where unpaid labor is not going to try to compete.

As some may have forgotten - or were not even alive when it happened - On July 20, 1969, Neil Armstrong, as part of Apollo 11 became the first human being to set foot on the moon. There are some interesting background issues which a number of people may not be aware of, and I'd like to share them.

1. Armstrong blew his line! When he stepped onto the surface of the Moon, Armstrong says, "It's one small step for man..." long pause, "One giant leap for mankind." I think the pause there meant that Armstrong realized he had not just blown it on national television, but on worldwide television. The line was supposed to be read as "It's one small step for a man, one giant leap for mankind." Rather than be caught with his pants down (hard to do when on the moon in 250 degrees F!), he finished the line anyway.
2. The transmission of the television signal used up almost the entire remaining spare national transmission capacity. My understanding was that basically, there was enough capacity in AT&T's microwave network to support 4 television channels. With the three networks using theirs (for programming being taped or filmed for broadcast later), that was essentially sll the extra capacity there was, one channel.

In a follow-up to the recent story here about how some Japanese schoolchildren will be tagged with RFID tags in their bags and clothing, CNN is reporting that the Mexican Attorney General (and some of his staff) have actually had microchip implants inserted beneath the skin, both as a means to access certain databases and as a means to be tracked in the event of kidnapping. Somehow this sort of thing bothers me, I fear it might become required.

A recent article here asked the question, "Favorite Programming Language Features?" and wants to know what feature or features you like most about any particular programming language. I'd like to ask the complimentary question: what do you think are the worst features of programming languages?

Here are some (of my opinionated) examples:

1. Case sensitivity in C. Of all the brain-dead moronic ways to make programming harder and more error prone, having case sensitivity has to be one of the biggest. It is difficult enough to write code and correctly use variables (unless your language supports mandatory pre-declaration) but adding the possibility of 2^N different identifiers for every identifier of size N (a two letter variable like "pi" has 2^2 variations (pi,Pi,pI,PI) increases the possibility of error.
   Now, however one of the nice features of at least one language is the use of case sensitivity with respect to formatting variables, in which the way the variable is declared causes the system to reformat every use of the variable to the same case (but you only have one variable of that name no matter what case you use to define it.) This allows you to determine if you have properly defined a variable, in that if you use the same convention and a variable's case doesn't change to match, you know you're using it incorrectly or it's not defined. In that sort of case, the language is making it easier for you to find errors, not making it easier to create them.
2. Pedantic demands for exactness in cases of harmless error. Most HTML browsers shouldn't mind if you leave off the </body> tags at the end of a document; presumably under XML if you don't close everything your document has errors (and presumably should be rejected as unprocessable). If you leave off a tag it should simply mean you get wrong results (like failing to close an <I> tag means that the rest of the document is rendered in italic or until a higher-level tag closes the range being processed, if the document's state is retained) from that point, not that the document is rejected.

Now, let's hear your comments.

In an article in ComputerWeekly, the FCC has ruled that it has final jurisdiction over unlicensed wireless space, meaning that an airport authority can't force airlines to (pay to) use its wireless network and may use their own. This bodes well for the development of wireless networks in various areas as it means that you have the right to set up your own network even if your landlord would want you to use theirs.

It was 20 years ago today, May 20, 1984, that The Terminator arrived from the future. It is interesting to consider how much of the world has changed in twenty years.

Warning: Plot Spoilers

For those who have never seen the movie, here is a summary: In the story, there has been a nuclear war, caused by machine intelligence seeing all humanity as a threat and triggering WW3 as a means to destroy that threat. The machines then operate death camps on a 24/7 basis to complete the extermination. A man named John Connor is able to lead the masses in an uprising to smash the machines and stop man's extinction. The Terminator is then sent back in time by the machines in an attempt to change its present by altering its past, specifically by killing Connor's mother before he is even conceived.

Having watched that movie many times in the past few weeks I am thinking how many things have changed in our world.

• Ubiquious reachability: while pagers were available then, they were relatively expensive and they weren't two-way. Today virtually anyone can afford a cell phone and anyone can get e-mail, even if you don't have a computer.
• Computers were very expensive then; today they are affordable by almost anyone, and the usability of them is tremendously better than it was when the DOS command line was king. Also today's desktop computers rival the power and capacity of mainframes then.
• Microwave ovens allow anyone instant hot tea or coffee or popcorn
• The legalizing of home taping spurred the acquisition of VCRs, which encouraged movies to be released on tape, which encouraged the development of DVDs
• The development of inexpensive fax machines has basically eliminated the need to mail short documents. In 1984, most newspaper help wanted ads for technical people had an address to mail resumes. By 1994, almost all ads used a fax number. Today, almost no place which is hiring anyone above a low-level peon job (and even some of those) lists a street address and either uses fax or e-mail.
• Unsolicited advertisements were basically received in the (postal) mail, which costs money. Today, spam represents 90% of the e-mail I get.
• Real-time communications across any but short distance was exorbitantly expensive. The Internet was unheard of and accessibility for ordinary people would be about a decade away. Today the Internet has collapsed space and distance for many people, has made many forms of long-distance communication either cheap or nearly free, and is a destructive threat to many organizations that collect rents for their middleman activities.
• Oh yes, a certain Austrian-born body builder was ripping people's hearts out of their chest in a popular movie. Today that same man is now the Governor of California.

So, what's your thoughts about how things have changed?