Older blog entries for nutella (starting at number 226)

Pump up the func
I've mentioned previously that I have been experimenting with some PBPK and enzymology modeling on a Ti89. This was my first exposure to anything that allows you to work digitally with symbolic mathematics. I must have been looking too awestruck as a colleague has allowed me to play with their installations of Mathematica (not seen it before) and MATLAB (only seen very old versions previously), just to expand my horizons. They are both very cool but have way too much functionality for my needs. At first glance the capabilities of Mathematica appear to be a superset of MATLAB (although the latter likely has the edge for matrix manipulation). I've started trying to move some of my models into the two packages. MATLAB looks as if it could handily replace some of my more clunky spreadsheets, where I'm using discrete methods, as I can turn the columns into lists and functions. Scanning Mathematica's abilities suggests that I could come up with more general analytical models, and that is very appealing. I see that there are some free alternatives but I don't know if any of them could cope with the uses I have in mind (I was aware of octave but have never used it).

Oh, the Good Old Days!
While browsing the shelves in the local library I came across a copy of the DVD from a concert I attended many moons ago. Naturally I had to borrow it and scan through it looking for myself (I'm still not a fan and so couldn't muster the time to watch the entire thing through). As I mentioned nearly seven years ago I was back behind the camera booms so my vanity wasn't tweaked. One question answered by the credits is the name of the mystery camerawoman - apparently it was Amy Ocheltree. Nice job, Amy!

It has been a while since I tried DOSBox so I gave it another whirl and am amazed at just how well it copes with some old programs. I may be able to retire my old 486/66 that I keep around "just in case".

...and a sample of blood. ...and your first born child.
My experience to date with CDW has been fairly positive, although I used to have the luxury of picking items up at the will-call window. Alas, recently they seem to have taken a sharp turn for the worse. I tried to order an item earlier this week and was told that, despite the web site claim of 2 - 3 days availability, it was out of stock for at least two weeks. I wasn't particularly surprised by this as I know that the availability numbers on the site are works of fiction intended to trap buyers. In the past I've had to wait a week or two for ~30% of the items I've bought. The latest shock was being told that, although my credit card details and address were okay, they could not sell the item to me as I had not given them my home telephone number. The only option to allow the sale to take place was for me to fax them an image of my driver's license. I checked with the sales person (who like all CDW people I've dealt with was friendly and polite). The summary of the email stream is as follows;
ME: Is this company policy?
CDW: Yes.
ME: Is there any alternative? You have validated my credit card and home address. Surely that's enough.
CDW: The only alternative is to cancel your order.
ME: What would you do with all my personal information?
CDW: My job is to collect all personal details from customers and if I can't do this with google, whitepages online or zabasearch I am obliged to ask the customer.
ME: Is it okay if I warn the rest of the world about this?
CDW: (implied) Yes.

So I can cross another supplier (and another vendor I'd endorse) off my list.

In addition to frequently falling into the well-recognised wikipedia trap and being distracted by reality I've noticed that I've now come to see inane email at work as slarshdawt articles and feel the urge to apply tags appropriately. Alas there's no way of making useful tags have an effect on those that most need them.

titus I sympathise with you concerning the changes in output format. I encourage you to keep going as it is very cool when you track down an apparent problem to an error in NIH's sequence file (or other supplier). The folk at the National Library of Medicine used to be very happy that other people were debugging their files by parsing them rigorously. I haven't done this in a few years so I am guessing that such source errors are rarer now but it feels good to be the one to spot them.

Thanks to Sun I now have the free (gratis) Solaris 10 DVD set. After sending my request in January I had given up hope as the intermediary who was mailing them emailed me to tell me that I hadn't filled out an application (so how did they get my contact information?). I've dabbled with Solaris 7 and 8 in the past, both x86 and SPARC, but the inclusion of Sun Studio on the latest set was what had sparked my interest. Hmmm... I have no idea if I can add an ATAPI DVD-ROM to my Ultra 5. Otherwise I'll try it out on a spare x86 box.

Get off my lawn!
It was nice to read about the 25th anniversary of the Sinclair Spectrum. The computer for my generation was actually the Spectrum's predecessor, the ZX81. Those were too expensive for me but I did manage to borrow my neighbour's (barebones 1 kB) box for a short while. Despite the limitations it was fun to program something in real time. The only alternative at that time was writing BASIC for an ICL 2900 mainframe (which I never actually saw) by sending the handwritten forms to be typed and receiving my output on 132 column green and white striped fanfold paper. The turnaround was about a week so proofreading in advance was a good idea. These days I am thoroughly spoiled as I can (potentially) debug my typos rapidly through live error messages. I was also reminded of those days recently when moving some of my modeling code to the Ti89. I have to use CellSheet, rather than a data file, and quickly ran into the 64 kB limit. I also became much better at proofreading my formulae (despite all the back and forth scrolling required) as some of the recalculations take longer than 5 minutes.

I was showing some of the modeling output (with Excel [sigh] rather than the Ti89) to my colleagues. There're two general reactions. The first group can see what I am up to and are fascinated that by using lab-determined enzymological and cellular parameters and basic biophysics I can predict more complex results. The other people stare at it and seem to be awaiting an Out of Cheese Error or the thing that goes "Boing!" or somesuch. "Yes, its a model but I'm only interested in real life"

For some reason this is the Scary Go Round that had me laughing loudest.

I only just noticed that I am certified as a Master on Advogato. I have no idea why (longevity?) and I definitely do not deserve it. Maybe badvogato has Advogato-juice.

My knowledge of statistics is pretty sketchy so it didn't come as much of a surprise to be told that the frequently used formula for variance of a sample (divide by n-1, rather than n for a population) is not as unbiased as I had thought. There's a fancy correction factor to take care of this and it uses the gamma function. Unless I am mistaken neither the basic Ti-89 arithmetic system, nor the official Statistics with List Editor flash application have access to the gamma function. Excel only has the incomplete gamma function as GAMMADIST() and, as I've mentioned before, the quality of Excel's statistical methods still sucks. Then I came across this wonderful page with many calculator-compatible approximations for gamma (and probably the clearest explanation of what the function is all about). Thank you Mr. Toth.

My brain hurts
One nice result from playing with the Ti89 is that long-comatose neurons are being prodded back into life when I need to perform some kind of mathematical workaround (my last formal maths class was a couple of decades ago). My current job entails analysing a lot of processes with exponential decays and most programs I use fit to the model;
y = A x e^Bx
For some reason the exponential regression on the Ti89 fits to the model;
y = C x D^x
I was puzzled for a while until digging through old maths notes reminded me that they were easy to interconvert;
B = ln(D)
I can now hide at the back and analyse data when meetings get boring...

Picks-up-the-mouse-between-pushes is now my insult of choice.

So, they really DO things differently in Texas...
As I mentioned earlier I am the recipient of a shiny new Ti89 and have been trying to get to know it. My main pocket calculator for [ahem] years, bought when I was studying for my M.Sc., was a Casio fx-3600P and I could use it almost blindfolded. If I had to make the leap from that one to the Ti89 I would be struggling as back in the days of yore calculators appeared to be more stack-based and so pressing <SIN> would give you the sine of the value on the display. Due to the LCD-leaking death of the fx-3600P I acquired a fx-115W with much the same capabilities but, being more than a decade newer, it now used what Casio modestly called Super Visually Perfect Algebraic Method which means you have to type everything like it appears in the textbook, or suffer painfully (thank goodness for some basic editing options). The Ti89 is like this, only more so. The main gotcha I have encountered so far is that unary minus has a lower priority than almost all other operations(!) Bizarre! When looking at a negative number I expect it to be atomic but the Ti89 treats the sign as some kind of function and it fails the dictum of least surprises as it is not followed by a parenthesis, like most of the other built-in functions. When you stare in wonderment after entering -3^2 and seeing the answer -9 you immediately have that sinking feeling that whole new vistas of misinterpretation and debugging are opening up. Is there any other language or machine that treats a unary minus in this manner?