Stupidity should be cured, says DNA co-discoverer. Okay, I don’t agree with him, because I think there are compelling arguments that we’re in over our heads when it comes to genetic manipulation, but I applaud him for speaking the unspeakable. And “People say it would be terrible if we made all girls pretty. I think it would be great.”? Gotta’ love it.
1. What is your favorite type of literature to read (magazine, newspaper, novels, nonfiction, poetry, etc.)?
I guess I’d say that of everything, my single favorite form is the extended essay — I actually read the New Yorker every week!
2. What is your favorite novel?
My stock answer for this is “The Princess Bride” by William Goldman.
3. Do you have a favorite poem? (Share it!)
I have a particular soft spot for “The Passionate Shepherd to His Love” by Christopher Marlow.
4. What is one thing you’ve always wanted to read, or wish you had more time to read?
Some of the great novelists have stumped me repeatedly with their long works (Dostoevsky, Joyce, Mann). And I wish that I had access to more plays.
5. What are you currently reading?
The Dante Club by Matthew Pearl.
Yesterday, I had a good meeting with Cory Linton and other members of Microsoft’s Tablet PC SDK. A Tablet SDK 1.5 is now available on MSDN. The major additions are a PenInputPanel which pops up a dedicated input window very near a normal WinForm control, which should allow for very fast porting / creation of form-centric applications, and an InkDivider object which essentially parses an arbitrary area of ink into paragraphs, text-lines, drawings and so forth.
If you’ve been reading this blog for any amount of time, you know that I think that mobility is the place for big productivity wins this decade. This is still ahead-of-the-curve stuff, meaning that now is an opportunity for bold and daring programmers to launch companies.
In case you missed it, here’s a 5-minute video demonstrating how easy it is to code the TabletPC starring yours truly. (It was my first experiment with a program called Visual Communicator — be kind.)
Borland Sidewinder to be “just” a C# IDE? Screenshots. Here are some screenshots of Borland’s Sidewinder tool for C#, which I thought was going to integrate TogetherSoft Control Center. No sign of that in these screenshots. Enough! I’ll actually…gasp… call them before posting further speculation. Stay tuned!
Open Source code better than commercial quality.
According to a study by Reasoning software (here’s a link, but it’s down at the moment), Linux’ 2.4.19 implementation of the TCP/IP stack ran around .1 defect per 1,000 lines of code (which we’ll assume means executable semi-colons), compared to 5 commercial OS’s (at least 2 of which were UNIX flavors), which averaged .6 to .7 defects per KLOC. Reasoning may only be looking at memory defects, which may explain why these rates are dramatically lower than the industry standard of 10-20 defects per KLOC (Software Assessments, Best Practices, and Benchmarks). The study is strong evidence for the OS argument that to many eyes, all defects are shallow.
1. What is your most prized material possession?
My wedding ring. (Awww…)
2. What item, that you currently own, have you had the longest?
Hmmm… I have the first release of the “Star Wars” soundtrack, (c) 1977. Since I only listened to it about three times (in 1977), it’s in near-mint condition. I wonder what I’d get on eBay for that…
3. Are you a packrat?
For books, yes. For anything else, no. Tina more than makes up for this, though, and most of our house is covered in bric-a-brac.
4. Do you prefer a spic-and-span clean house? Or is some clutter necessary to avoid the appearance of a museum?
5. Do the rooms in your house have a theme? Or is it a mixture of knick-knacks here and there?
It’s a mixture of travel souvenirs, original art pieces (mostly oils and ceramics), books, and computers.
There are three ‘platforms’ for building corporate applications today: J2EE, Microsoft.NET and the ‘Open Source’ platform. In the later I also include Java software that is not J2EE like Struts, WebWork, Hibernate, Castor….a big oil company and an airline company. Both have J2EE as their standard platform….Slow moving is not a problem for most corporations. They usually move slowly, and they feel comfortable with ‘designed by committee’ technologies, so J2EE has a good value proposition for them….
via [Andres Aguiar’s Weblog]
Andres does a good job articulating one aspect of a fundamental tension in corporate software development — the processes by which software development is done today are known to be inefficient and expensive, but they’re known and therefore somewhat controllable. Many corporations would prefer to budget $100 believing it’s +- $10 rather than $80 that’s +- $25. In no small part this is because for 40 years, they’ve been told that new processes “cut development costs in half” or even “by a factor of 10.” I estimate the overall productivity advantage of .NET versus J2EE above 10% and less than 20%, which is tremendous, but a MS reviewer criticized comments to that effect in Thinking in C# as “overly cynical and detracts from the author being viewed as the authority on the subject.”
These are the numbers that Microsoft uses in their own VS.NET 2003 ROI calculator: Unified IDE +10% productivity. Drag-and-drop server components: +10% productivity. OOP: +10% productivity. Debugger improvements: -20% debugging time. Improved test scripting: +8% test productivity. Web forms designer: +25% Web UI productivity. Dynamic help: 24 hours per developer per year. Even if you accept these numbers, (–25% debugging time? Hah! Advanced debugger features have minimal impact on the overall time spent correcting defects!), how in the world can you argue the tools give you >20% productivity over J2EE using, say, IDEA and jUnit? It’s just not true.
Microsoft has filed broad patents covering the .NET API. It appears that they are attempting to patent the relationships between namespaces (client programming versus XML manipulation versus network transport, etc.). At this broad level, the patent should be rejected on the basis that the particular separation of concerns noted is obvious “to one of ordinary skills in the art.” Reading the entire patent, though, although the claims relate to very obvious high-level relationships, the supporting documents mention every single namespace in the .NET Framework — as if the specific patent is for the admittedly non-obvious web of relationships between all the namespaces, a claim that should be rejected on the basis of operativeness. Operativeness is something like a negative proof — if an invention that does not implement the exact structure is useful to the extent of the claims, the invention is not patentable. For example, you can patent a new design for fastening sheets of paper, but you can’t patent an obvious paperclip design in which the paperclips are blue (unless you specifically make claims about the usefulness of blue-ness).
This is a bad patent and should be rejected. Let’s see if the $1B per year spent on patent review does the job this time…
Yes…. my first personal computer was this close (fingers teeny bit apart) to being an Amiga. Instead I bought an XT-compatible with Hercules graphics.