Wednesday, February 27, 2008

Recalling the Commodore 64


Tech Republic today (Feb. 27) ran an interesting historical story "Cracking open the Commodore 64," by Mark Kaehlin, link here (with pictures -- may require subscription). The computer, introduced around 1983, could not only play computer games but could connect with a community online at 300 baud. (That was the year the movie "War Games" came out.) By 1989, some companies still thought of 1200 baud or 2400 baud as an acceptable speed for a connection to a small IBM mainframe from a remote location (9600 baud was a real luxury).

My first computer was a Radio Shack TRS-80 at the end of 1981, with its 64 character black and white monitor, a unit that cost $3700. Everyone was talking about the Osborne, the Atari and, yes, the Commodore; the IBM PC (with "PC-DOS" instead of "MS-DOS") then was a real "luxury."

Friday, February 22, 2008

Dell's "system voltage low" message


For the first time in four years, I got a “system voltage low” message on my Dell 8300 computer as I cold-booted it. It said that I could continue with F1, which I did, and the system booted and connected to the Internet without incident.

I don’t recall whether the word “battery” occurred in the message. I looked around, and I see that Dell can give that message when the CMOS battery is failing, or when there is not enough voltage from the surge protector. I have a Belkin UPS.

The system date and time were correct, and had not gotten out of sync. I booted two other times this morning without the message. I wonder if this happens when the line voltage outside is too low, which could have been the case since there was a mild ice storm in the area this morning, but no outright power outages right here.

I found a couple or references, on the Dell 8300 here.

And another on other Dell machines.

I wonder if this is a quirk in how Dell is engineered. Literature says that CMOS batteries should last about ten years. This is four now. It looks like they sell for about $30. I don’t know whether whole 115-volt power supplies fail, but it’s probably rare.

In 1997, I bought a Compaq laptop from BestBuy in Minnesota, and the first two items had power-supply failures almost immediately. The third machine worked and I still have it ten years later.



Dell error messages: link.

Another link on the system battery voltage: seems to say bypass surge protector and connect to wall outlet. Bizarre. link.

Or this.

Best answer. It seems some CMOS batteries stop recharging. Look here.

Dell's reference for 8300 link.

Dell's replacement instructions are here.

Hardware tech explains why CMOS doesn't recharge here.




Update March 21. Possibly associated with surge protector or UPS device after all


The message I get really says "system battery voltage low." The latest reference that I could find on Dell is here. The date and time from CMOS do not stop working. The visitor may need a service tag code to use this link, and the website will generate a journal reference number for your visit. The link indicates that the problem can be caused thy the way a surge protector is grounded, or by certain Belkin products, and suggests that an event log can be cleared. It is not clear that the Belkin (which I have) is not delivering enough voltage, but the computer operates absolutely normally otherwise. This contradicts other widespread reports about CMOS battery failure causing this message. If anyone else knows something, please comment.

Tuesday, February 19, 2008

Career Coach appears on Nightline: how to sell yourself


ABC "Nightline" tonight had a session with career coach Peggy Klauss, who did a sample "career rescue" session with a laid-off 40-something data analyst (and quality assurance)person in Tucson, Arizona, at the man's home.

The ABC story is "Learn to 'Recession-Proof' Your Career Career Coach Peggy Klaus Offers Key Strategies to Survive an Economic Downturn," link here.

Her program consists of steps or "lessons." For example, "The workplace is no place for modesty."

I remember this idea from Barbara Ehrenreich's book "Bait and Switch", about how important enthusiasm is to employers. It does sound like an awful lot of self-salesmanship is about manipulating words (sometimes with hyperboles) and impressions. In fact, her client, after his moderate makeover (and extreme makeover of his resume) started getting calls from recruiters immediately.

There was a tendency in the 80s and 90s, through Y2K, to be the "super prole" who could do the job hands on, and who could jump from one task to the next. That seems to have changed since the 2001 recession. Now, specific expertise is more important. So is advancement, but it should be advancement related so self-chosen goals. And so is passion. Perhaps young Clark Kent in the Smallville series has the right word: "destiny." That may not be COBOL anymore, or even C##. Or it make be something arcane. It may be what you believe you were "born" to do after all.

Friday, February 08, 2008

Short training courses in OOP's don't help a lot


As I mentioned in the last post, back in the early 80s, companies often conducted training by “VTR” lessons in programming languages, databases like IMS, CICS, often supplemented by audio tapes and written quizzes.

In the 90s, in office parks around all major cities, companies proliferated to offer training in all the hot new languages, most of all Java, which came into being around 1996 but became accepted in production end-user data access systems amazingly quickly.

Typically, a course would be one week, five eight hour days, with many breaks, plenty of coffee-cups and doughnuts (topology pun intended) and things that cause sleepiness and hypoglycemia. There would be plenty of classwork exercises, but not really enough time to do them. Most of the time students who took these had already been exposed to the languages at work – especially for Java and Powerbuilder. For someone with mainframe experience only, say around 1999 or 2000, the pace became fast, and many of the classwork assignments involved being able to figure out how to do something from the “Help.” Because the construction of classes and objects is non-sequential activity, this is hard to get out of the blue. Another good example of a class like this was BEA webserver in 2001. Classes like this always cover a lot of material that the typical analyst may not need in his own job.

Community colleges offer training in pretty much all of these topics. In the fall of 2002, I took courses in XML and C#. The XML course had regular quizzes and a project, which I actually integrated with material from the C# course, with Visual Studio .NET. The C# course grade came entirely from homework. Every week there would be assigned problems, and it turned out that in most cases there were several approaches that could produce acceptable “solutions.” The instructor always had more compact approaches, however, than a student would.

The only way people get good at object oriented program is to “do it,” and get into a project early, code a lot of classes and methods, QA and implement them. Object oriented programming is very difficult to “get” from a course or textbook alone. The knowledge and skill required by Sun for java certifications is very great, and unachievable without a long period of hands on experience. University IT curricula have to be designed to give the students enough volume of hands on experience in a variety of contemporary problems. It’s amazing, however, how quickly Java came to the front of the pack.

Monday, February 04, 2008

From coding sheets to punched cards, to "tube cities", to VTR's, VHS and Betamax -- to DRM today


There were days of paper and pencil even in information technology. I remember well the columned coding sheets, with the columns in zones according to the rules of COBOL, FORTRAN, and Assembler, all the way back to my first summer job (FORTRAN) at the Naval Ship Research and Development Center in 1965. At Bradford, in1977 and 1978, we still coded new programs on sheets and sent them out to little companies in midtown Manhattan to be keypunched onto Hollerith cards. We would edit them online with Roscoe and ISPF, but there would be only about four terminals for fifteen programmers. (At NBC, with a Univac 1110 system in midtown, everyone had a teletype terminal and use it; Univac in those days was ahead of IBM in programmer-friendly technology even though keypunch sales had been a big business for Sperry Univac in the early 1970s).

We also wrote program specs in longhand, often on forms, and gave them to “typing pools.” In 1979, a Lanier “word processing” machine cost $15000. Even when I got to Chilton Corporation in Dallas in 1981, programmers walked to a “tube city” to edit code “online”.

At that time, around 1982, my project leader wrote “PDL’s” – “program descriptor language” – essentially pseudocode – on paper forms and assembled them into large black three-ring binders and handed them to management for review. These “programming specs” for a new daily billing system amounted to perhaps 200 pages of these forms. Management discouraged his making a Xerox copy (“wasteful”), so when he turned in his work, he had no record of exactly what he had written. I started the practice of writing all my specs on Roscoe.

At the time, we were replacing a system (with COBOL, MVS, ADR DATACOM DB/DC) that ran on an Amdahl but written in DOS assembler executed under “DUO.” A lot of smaller shops then had just DOS.

I recall when starting to work there that I took about two weeks of afternoon training in various subjects, starting with “structured systems design,” with each lesson comprising a VTR video (they called them “video tape recorders” then, and the terminology for VHS would soon come about), and audio tape, and a paper quiz. In those days, tiny little cassette recorders for meetings were all the rage. It’s interesting to ponder the VHS Betamax case in 1984, that would help the stage for litigation that is so critical for digital rights management today: namely whether manufacturers could be held liable (“downstream liability”) for creating technology that consumers could use for infringing purposes. The EFF reference is here.

How times have changed in three decades. (I actually had a Beta player in the 80s. Wrong choice.)