photo-rbt.jpg (2942 bytes)

Email Robert

Daynotes Journal

Week of 23 August 1999

Sunday, 29 August 1999 09:58

A (mostly) daily journal of the trials, tribulations, and random observations of Robert Bruce Thompson, a writer of computer books.


 

 

 

Search this Site

TTG Home

Robert Home

Daynotes Home

Special Reports

Current Topics


Monday, 23 August 1999

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]


I spent some time yesterday, gasp, cleaning up my office. There's a lot still to be done, but now at least I can see the surface of my desk. I even found some stuff that I'd forgotten I had. I'll continue cleaning up over the next couple of weeks, until I eventually get Organized. The problem is, I'm not naturally a neat person. So I'm likely to have a continuing cycle from neat to cluttered. My motto: A Place for Nothing, and Nothing in its Place...

* * * * *

I stumbled across yet another computerish daily journal site the other day, this one kept by Chris Ward-Johnson, aka Dr. Keyboard, the computer columnist for The Times. His daily journal is interesting, funny, and well worth reading. You can also read Chris's current Times column here. I started reading his daily journal with the first entry back in May, 1998 and have gotten through a couple of months so far. 

* * * * *

This from Mitch Armistead [mhaathome@worldnet.att.net]: 

I was playing with Wins the other day, and configured a NT machine not to be a Wins client. It was my understanding that a Microsoft Hybrid Node would ask Wins, then do a broadcast. No machine on the same subnet (all H Nodes) could see the non Wins client.

Is there something obvious I'm missing?

Any help greatly appreciated.

Probably not. WINS is a hateful service, a kludge that Microsoft implemented to graft its obsolete NetBIOS networking schema onto IP. When a WINS client machine starts, it registers its NetBIOS computer name and IP address with the WINS server, making that information available for lookup by other WINS clients. Your non-WINS client depends exclusively on broadcasts to resolve computer names elsewhere on the subnet (unless you've created an LMHOSTS file that maps NetBIOS computer names to IP addresses). Your WINS clients query the WINS server to locate other machines unless you explicitly do something that causes them to broadcast. For example, if you choose Start - Find - Computer on one of the WINS clients and give the name of the non-WINS computer, it fails to find that computer in the WINS database and so does a broadcast that basically says, "if your computer name is such-and-such, will you please send your IP address to me?" If you use Start - Find - Computer, I'll be you'll be able to find your non-WINS client.

* * * * *

This from Joshua D. Boyd [jdboyd@cs.millersv.edu]: 

About displaying computer video on a TV screen. Have you considered replacing your TV set with a Gateway Destination monitor? They are big screen (I think 37" is one of the possible sizes), low res monitors (800x600 tops, I think). The other possibility is to just use a video card that has a TV out jack. The upside is keeping your old TV, the downside is that TVs aren't meant for displaying anything but moving video, so when you are in windows (using non full screen programs) it is going to look awful. And the technical resolution of NTSC television is something like 768x512, although frequently some of that gets clipped off the sides. Another option which is pricey (but then, that 50gb drive is also, so I thought I present it) is to switch from a TV to a projector. Someday that is what I plan to do. I'm thinking that a projector that can do 1024x768 would be good, shined against a full wall. That way I can have a really large video window, but also keep other things open it I want to (like instant messaging, or stock quotes, or what ever i'm into then).

As far as keyboards go, I have one that I love. It is the size of a notebook keyboard, and it has a built in touch pad. There is a wireless model available. This keyboard has no brand name on it anywhere. Mine cost $39, and the wireless version cost $49. I bought it mainly for the reduced key travel (approx. 2mm instead of the more normal 4mm.), which reduces RSI related stress, and I'm finding it is much quieter than full sized clicky keyboards (clicky being used here to denote keyboards that don't use those bubble switch things).

What exactly do you plan to do with this machine anyway, especially with all that HD space. If you were a student, I'd guess that you were going to download tons of MP3s and movies before they come out on tape. But even for that so much HD space is over kill, since CD-ROMS are much cheaper (granted, you usually have to switch disks halfway through the movie). If it were me, I'd being doing something like acquiring the Visible Human Dataset (15 gigs for the man, 39 for the woman), and/or using it to hold rendered frames of video (currently to takes a lot of effort to hold the video before I can compress it to mpeg).

--

Joshua Boyd
http://catpro.dragonfire.net/joshua

I'll probably stick with the existing TV and a use a video card with a TV-out jack. This system isn't intended to be a general-purpose computer, or I would go with one of the large monitors. All I really need computer video-out for is managing the a/v server itself.

As far as what I plan to use it for, I'm not sure yet. In general, storing audio and video files and making them available locally (to the TV and audio rack) and across the network. I've downloaded a few MP3s from the net, but I'm not really interested in that, not least because they're typically compressed at 128 Kb/s or less. I plan to rip our existing collection of CDs, probably at 256 Kb/s or more, and store them on the server, where they'll be accessible from anywhere in the house.

Another thing I'd like to do is use the server as a kind of digital VCR. That is, if a movie is coming on AMC that I want to tape, I'd set the AV server to record it to disk as MPEG-2. And that raises another issue. I think I know that MPEG-2 recording is non-trivial. Even decompressing MPEG-2 data takes a dedicated decoder card if you want it done properly, with high quality and no dropped frames. I can't imagine how much more horsepower is needed to compress MPEG-2 in real-time, but I'd guess that it must be substantial. So how do I take that inbound analog data stream from AMC or whatever and compress it in real-time as MPEG-2?

Then there's recordable DVD. Right now, a DVD-RAM cartridge is about $20 to $25, but I'd imagine that the price of cartridges will come down fast if DVD-RAM wins the standards war. If that happens, I'll be able to store movies on a DVD-RAM cartridge for not much more than a VHS tape costs. The problem with DVD-RAM right now is that it's limited to about 2.6 GB per side, which translates to about 80 minutes of video, not enough to tape a move. But Hitachi tells me that their next-generation DVD-RAM drives, due out this fall, will store more than 4 GB.

We'll see what happens. Right now, this is a project system in the "thinking-about" stages.

* * * * *

This from Bo Leuf [bo@leuf.com]: 

Interesting figure in the news:

"The number of U.S. adults behind bars or under police supervision last year reached a record 5.9 million offenders, almost 3 percent of the adult population, the Justice Department reported Sunday."

However, just as interesting was the statistic (Newsweek) that despite the recent "surge" of high-profile shootings, esp in schools, the total number of violent "incidents" in US schools had continued its sharp decline from about 92 and was again markedly lower this year.

/ Bo

--

"Bo Leuf" <bo@leuf.com>
Leuf fc3 Consultancy
http://www.leuf.com/

Yes, and the really interesting thing is that the majority of those people are there because of this insane "war on drugs." We have hundreds of thousands of people incarcerated who did nothing worse than get caught with relatively small amounts of cocaine. We've built new prisons to hold them all. If you factor out all those people who have been jailed under our Draconian drug laws, you find that US incarceration rates are similar to other developed countries.

* * * * *

This from Jan Swijsen [qjsw@oce.nl]: 

Exactly. And for nearly any other information handling endeavor, computers have been of equal benefit. And as for why you're getting more work done and overall productivity isn't increasing, it's because those who would otherwise be doing useful work are in fact doing government-mandated busy-work instead, whether on or off the government payroll.

Of course in America the government is the cause off all problems :~)

My father writes school books for Geography. He has been teacher (paid by the government;-) ) until his pension and has been writing books for about 15 years now. Originally he did send the text printed out in double spaced text. It came back on large test print sheets which had to be cut and pasted (by hand) to insert photos and graphics. All this was taped together and returned to the publisher. It came back on test print sheets in the expected layout for commenting and changing. The graphics were done by crayons and then a graphics artist did them, again by hand, in colour separations.
Typically, during these editing stages several people from the publishers were working full time on the book. These were typically the same people that worked on the previous books or other similar publications.

Today he is finishing a book again. For him the whole process takes at least as much time. Only he has to do all the layout himself, there is no proof reader anymore, there is no dedicated graphics artist. The graphics scanning and converting is not done by people from the publisher but the work is out-sourced. The people doing it don't know Geography and are not long enough on the job to understand the specific issues and thus maps and graphics have to be checked better and have to be redone often. The same with layout often the wrong fonts or point sizes are substituted causing the text to go over more than the allowed page number.

The productivity gain of using computers here is mainly swallowed by having people spend time on converting files and fiddling with settings. And by putting inexperienced people on jobs that require a certain level of experience.

No doubt the government also steals from the possible productivity gain but it is by no means the sole or even greatest black hole. (Here in Belgium government regulations are slowly becoming more streamlined, laws are actually being replaced iso being amended. Still a verrrrry long way to go though.) The greatest problem, for productivity, is that as more things become possible more things will be done wrong and will require correcting.
Quick question : how much of your time is eaten by computer related trivia (not related to your job not directly to the government) ?

ps
Removing Rudzki's Rant-mailings from your daynotes clears the air considerably. He keeps on ranting but on his own site as he should. He rants about his neighbours causing him discomfort. Quite rightly so. But then he complains because someone is discomforted by him and says so. Not very consistent.

Kind regards,
Svenson

Well, I don't maintain that government is the *only* reason that productivity is stagnant, just most of the reason. In economic theory, productivity is determined by the ratio of output to input. By that definition, the book your father produces today requires many fewer resources than the book he produced years ago. That equals higher productivity. The fact that the additional resources that used to be available to him are no longer available doesn't change the fact that your father's productivity is now higher due to his using a computer. He's accomplishing more work with the same amount of effort. The question is, why are those additional resources no longer available? 

In a global sense, it's because useless government rules and regulations have directly or indirectly claimed resources that would formerly have been available. 

Directly, in the sense that the publishing company may now have a smaller "teeth to tail" ratio because a higher proportion of its employees are now required to deal with government regulations instead of doing the work of the business (e.g. twenty or thirty years ago, a given company might have a "personnel" department with 5 employees; today, that same company has a "human resources" department with 20 employees). 

Indirectly, in the sense that ever increasing taxes reduce the size of the pool that employers have available for hiring and paying people to do the company's work. Those resources subsumed by the government are used to pay an ever increasing number of government employees to do useless work. Simply examine the total number of government employees as a percentage of all employees, or figure the number of government employees per thousand population. Do that for any industrialized country and I think you'll find that the percentage of government employees is much higher now that it was ten and twenty years ago.

 


 

 

 

Search this Site

TTG Home

Robert Home

Daynotes Home

Special Reports

Current Topics


Tuesday, 24 August 1999

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]


Trying to finish up a chapter today and tomorrow in preparation for leaving Thursday for Asheville, NC, where Barbara will be attending a one-day conference. There probably won't be much here later in the week. Come to that, there's not much here today other than letters.

* * * * *

This from Chuck Waggoner [waggoner at gis dot net]:

It really baffles me that Intel is going to such extremes to prevent overclocking--now an article in THE REGISTER describes a new technique that has apparently got Microsoft cooperating by modifying NT4 SP5 so that certain CPU ID's are impossible to overclock.  And the person who found this, was using regular Pentium II's and III's--he wasn't trying to turn a cheap Celeron into a Pentium III.  Can the management people at Intel who are apparently so overly dedicated to killing overclocking really be playing with a full deck?

If somebody fries a chip by overclocking, so much the better for Intel--they've sold another CPU to replace the burnt one.  And those who will buy the extra hardware to overclock, can't be more percentage-wise, than those guys who used to 'soup-up' a car with carburetor, muffler, timing, and gear enhancements that squeezed extra torque and horsepower out of a factory drive train back when we were kids.

Intel should be PROUD their CPU's are of such high quality that overclocking is possible.  The car companies used to tout how well-suited their cars were to aftermarket enhancements.  Instead, Intel asserts a public image that at the same time resembles a big-brother ogre and a little kid who wants to make all the rules or takes his toys home.

My father once worked in sales for a company that manufactured piston rings.  The company had a great story about how they used negative salesmanship to get one of their largest orders ever.

This company manufactured their piston rings in a manner which, when properly fitted and installed, had a natural "spring" in it that made a tighter seal than any of their competitors.  Installation, however, required a different method than filing the gaps by hand, which was the common method at the time.

One particular mechanic, who had been trained in the new installation method by his auto parts supplier, nevertheless continued to file the newer "springed" rings.  This destroyed the ring's ability to make the tighter seal, and left it no better than all the competitors' product.

Finally, a District Manager for the manufacturing company went to the mechanic and tried to explain to him why filing was a bad procedure for his customers.  The mechanic replied that he had been installing rings for years the same way and he intended to keep on doing it.

So the District Manager said to him, "Look, if you're going to keep on doing that, will you please use some other brand of piston rings.  We do not want our rings installed in that manner."  He then he left the mechanic.

Next day, the mechanic went to the parts supplier with a big box and reported what the District Manager had told him.  Then he said, "I'll use what piston rings I damn well please, and that guy can go to hell!"  He then bought the largest supply of those piston rings that had ever been placed by an individual mechanic.

And the home office was well satisfied with the outcome.

--Chuck Waggoner [waggoner at gis dot net]

I saw that article yesterday as well. I thought about mentioning it, but I hadn't verified it and The Register didn't seem any too sure that it was true, so I didn't. As far as Intel and overclocking, I'm not sure what's going on. Historically, they haven't much cared about individuals overclocking their CPUs. What they do care about is dishonest companies remarking their CPUs, which was the real reason they implemented the multiplier lock. Now that the only way to overclock is to run a 66 MHz FSB CPU at 100 MHz, there's not that much opportunity for remarking, so I'm not sure what this SP5 thing means, if indeed it's true.

* * * * *

This from Joshua D. Boyd [jdboyd@cs.millersv.edu]: 

I'll probably stick with the existing TV and a use a video card with a TV-out jack. This system isn't intended to be a general-purpose computer, or I would go with one of the large monitors. All I really need computer video-out for is managing the a/v server itself.

As far as what I plan to use it for, I'm not sure yet. In general, storing audio and video files and making them available locally (to the TV and audio rack) and across the network. I've downloaded a few MP3s from the net, but I'm not really interested in that, not least because they're typically compressed at 128 Kb/s or less. I plan to rip our existing collection of CDs, probably at 256 Kb/s or more, and store them on the server, where they'll be accessible from anywhere in the house.

As I have time I'm working on encoding my CD collection as 256k files. I'm starting to run low on disk space though, so I probably won't get to much farther until I scrape together the funds for a larger hard drive.

Another thing I'd like to do is use the server as a kind of digital VCR. That is, if a movie is coming on AMC that I want to tape, I'd set the AV server to record it to disk as MPEG-2. And that raises another issue. I think I know that MPEG-2 recording is non-trivial. Even decompressing MPEG-2 data takes a dedicated decoder card if you want it done properly, with high quality and no dropped frames. I can't imagine how much more horsepower is needed to compress MPEG-2 in real-time, but I'd guess that it must be substantial. So how do I take that inbound analog data stream from AMC or whatever and compress it in real-time as MPEG-2?

Normal CPUs can't compress mpeg2 in real time. You will need an encoder card for that. But most video capture cards (except a few really high end ones) will have acceptable compression systems, even if they aren't mpeg 2. I can't really make any recommendations though. Are there even Windows programs to do what you want? I've heard that doing such task under windows 9x is highly unstable (so don't get rid of your VCR), and you could do that stuff under NT, but NT doesn't support DVD yet. 

Then there's recordable DVD. Right now, a DVD-RAM cartridge is about $20 to $25, but I'd imagine that the price of cartridges will come down fast if DVD-RAM wins the standards war. If that happens, I'll be able to store movies on a DVD-RAM cartridge for not much more than a VHS tape costs. The problem with DVD-RAM right now is that it's limited to about 2.6 GB per side, which translates to about 80 minutes of video, not enough to tape a move. But Hitachi tells me that their next-generation DVD-RAM drives, due out this fall, will store more than 4 GB.

That will be interesting to see what happens there. I'm looking forward to when I will be able to build my own media server. In my case it will most likely be a linux machine, although I might go with BeOS as well. It depends on where BeOS is by then (then definitely not being anytime in the next year).

We'll see what happens. Right now, this is a project system in the "thinking-about" stages.

I understand how that goes. I have the parts to build myself a dedicated MP3 server, (if I actually use it much it's going to need some upgrades, but I can at least build a concept machine) and I've been planning it for a long time, but I've never gotten around to actually doing it. I'm not really looking forward to making linux work with the no-name sound card I have for that computer, and worse, it just isn't a pressing need.

--

Joshua Boyd
http://catpro.dragonfire.net/joshua

This is definitely a blue-sky project at the moment. As you say, the software may not even be available. Not only do I not know all the answers at the moment, I don't even know all the questions. I'd like to stick with MPEG-2 rather than a proprietary compression system simply because I want those video files to be readable and playable by other systems on the network. But we'll see what happens.

* * * * *

This from Jan Swijsen [qjsw@oce.nl]: 

"(e.g. twenty or thirty years ago, a given company might have a "personnel" department with 5 employees; today, that same company has a "human resources" department with 20 employees)."

What I wanted to point out, and your example comes in handy here, is slightly different from your view.

The publisher was one big company, it had perhaps 10 people in "personnel" department.

Now it is much smaller and has only 5 "human resource" "managers".
But it subcontracts out work that it used to do itself. Now there are about three or four small external companies involved in the process of producing the book. Each of these adds one of two additional "human resource" people making up the loss in the old company.

The result is no change in productivity, just a shift.

There is also an added overhead of communication between the companies that eats productivity improvements in some other part of the process.

And there is the government that takes its "fair share" of the productivity gains.

"...productivity is determined by the ratio of output to input. By that definition, the book your father produces today requires many fewer resources than the book he produced years ago."

If you cut out the government (nice thought) you still don't reap anywhere near the possible gains. For my father it means that he has to work much harder, doing the job that others were doing before. Most of the gains he gets from using computers are eaten by extra companies involved. He invests more of his time, doing more of the work (more input) while more companies are making a small profit. He doesn't get more money for it (no increase in output). So there is no productivity increase for him.

The combined profit of the companies should be higher but it is here that the government cuts in again.

Right, but there are two separate issues here. There is no doubt that your father's productivity has increased, i.e. that by using a computer he is able to produce more output than he did before he used a computer. That obviously frees up the resources that used to be required to do the additional work that he is now doing, allowing those resources to be used to produce additional output. In the natural course of things, your father's additional productivity would mean that he could also consume more, i.e. that he would be paid more for that additional output. The fact that he is not means in a macro sense that that additional productivity potential is being sunk somewhere. My point is that it is being sunk into wasted, government-mandated busy work that results in no additional goods or services. That is, if everyone is twice as productive now as formerly, that means that there should be twice as many goods and services available to consumers, and everyone should be able to purchase twice as many goods and services with the earnings from their own output. The fact that they cannot means that that surplus is being sunk somewhere. Guess where.

 


 

 

 

Search this Site

TTG Home

Robert Home

Daynotes Home

Special Reports

Current Topics


Wednesday, 25 August 1999

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]


Something that's beginning to concern me a great deal is the unpredictability of Microsoft software. And I don't mean simple bugs. All software has bugs. In the past, though, those bugs have been predictable. If you do A, then B, then C, a Bad Thing happens. If you again do A, then B, then C, exactly the same Bad Thing happens. That still happens with current Microsoft software, of course, but that's not the kind of bug that really worries me.

It seems that recent Microsoft software increasingly behaves unpredictably in response to the same series of inputs. For example, I've mentioned before that the font size used by Internet Explorer changes at random. I've set IE to use the Larger font size. But when I fire up IE, it comes up half the time in Larger and half the time in Medium. This with no other programs running. If I'm viewing a page using Larger and open a new instance of IE, it may come up in Larger or it may come up in Medium. I can, with nothing else running, fire up IE, watch it come up in Larger, close it, fire up IE again, and have it come up in Medium.

Same thing with FrontPage 2000. Every morning I do basically the same thing. Update my current page (right now, /daynotes/0823RTDN.html), save it, copy it to the /rbt folder, and rename it to /thisweek.html before publishing. Sometimes, FP2K prompts me whether or not to overwrite the older version of one or the other (or both) pages on the server. Other times, it simply publishes the new pages without prompting. This morning, it prompted whether or not to overwrite the old version, which it said had last been modified yesterday by "Unknown". That's pretty strange, given that I'm the only one who edits these pages, and it usually asks if I want to overwrite the page modified on such-and-such a date by "thompson".

Another example. I was downloading several driver updates the other day, using IE's built-in FTP function. The first time I told it to download a file, it brought up my profiles directory on C: as the default download location. I changed that to point to f:\install\intel and started the download. Then, in the same instance of IE, I hit another web site and started another download. This time, IE defaulted to storing the downloaded file in f:\install\intel, which makes sense. I changed that target folder to f:\install\matrox and proceeded to download the Matrox. Then, again in the same instance of IE, I hit yet another web site and started another download. IE again defaulted to the last-used download directory, f:\install\matrox. Fine. I changed that to f:\install\adaptec. I started yet another download from yet another web site. This time, instead of defaulting to f:\install\adaptec, IE suddenly decided to default to storing the file in my profiles directory on C: This ain't good.

This trend is very disturbing to me. If I provide a given sequence of inputs and the Microsoft software outputs "apples", it should do the same every time. Not "apples" one time and "oranges" the next. Reproducible bugs I can live with. But these weird bugs are starting to get to me.

* * * * *

This from Wallbridge, Shawn [shawnw@elections.mb.ca]: 

Here is a little blurb I found at http://www.hardocp.com/. This was posted a few days ago. The update is the funny part.

Although you included the text of the article, I decided not to print it, because it is copyrighted. Instead, I went to the hardocp page itself, hoping there'd be a link for it. There's not, but the article was time-stamped August 23, 1999 -12:45pm for anyone who wants to read it. I agree that it's worth reading. 

* * * * *

This from Jan Swijsen [qjsw@oce.nl]: 

That is, if everyone is twice as productive now as formerly, that means

...

This sentence is the clincher.

If everybody is twice as productive, but our total output is the same as it was, then Government must be twice as big as it was.

Government (in Belgium) has not doubled since the start of the PC age. Grown certainly (well I don't have any figures but I assume it did) but not doubled. My conclusion, not everyone has become twice as productive.

As I mentioned in my original mail a lot of the work my father does now was done by experienced people at the publisher. Now he does them himself because the quality from the publisher (subcontracted to...) is unacceptable. The service is still available but using it resulted in so many returns that it became faster do do them himself.

For example we had a blind map of Poland scanned in mirrored. No problem everybody can make a mistake. Only it was supplied with overlay sheets for colour separation and those we scanned in correctly. Resulting in colours that absolutely don't match with borders. Another example, he (and the co-authors) use WordPerfect v6.1 on recommendation from the publisher because they can read that file format on their Macs. They also recommended using Helvetica as type font. Now half the time print samples come back with the wrong fonts or font sizes. When asked they claim it is because they had problems with converting the files. Very convincing. These are just two examples but I could give you more. (I could pass you my father but then, before you get him to stop ranting, you impression of Belgians would probably sink too low.)

I don't think that things like this can be attributed to the government.

This type of sloppy work is becoming increasingly common because people are more looking to economic 'opportunities' than to stability and quality. Building a quality reputation is a long time job, it is hard work and it doesn't produce a high return on investment. At least not in the short term that investors are interested in.

And of course specialization adds in as well. Answers like 'I am a graphics specialist, what do I (have to) know about geography' are seen as acceptable these days. And not much can be done because the service is outsourced and going to another bureau hardly ever improves the situation. Often the founder of a bureau is the only one that can do the job right but cannot do it because he has to spend his time on customer relations (to get orders and to get a higher financial turn over) with too little time supervising and teaching his newly hired staff. (You would argue that also spends more time doing government mandated bureaucracy than anything else. That however is outsourced and has been for years so the service quality is reasonable.)

Government has an impact on this (of course). By making it easier to start a new business it has, since a few years, become profitable to outsource such specialized jobs. What is a good measure from the government (for once) is causing problems now. Probably quality will surface in the long run but the current financial climate doesn't favour long term prospects. Not while 'next year' is long term.

Regards,

Svenson

ps is it just an impression or am I really ranting here?

Well, I don't know enough about the government and business environment in Belgium to comment, but here in the U.S. we have many of the same problems, and many of those are due to government interference. For example, Microsoft has recently been taken to task for the large numbers of contract employees they use and how they treat those employees. But the reason that Microsoft is using contract employees in the first place is government rules and regulations. If they could hire and fire at will, they'd simply employee these people rather than going through the convoluted contract employee process. Although the US has not yet gotten to the state of some European countries, where it is nearly impossible to get rid of an employee once you hire him, it is now simpler in many cases here to simply contract out work to individuals or companies rather than hiring the in-house expertise necessary to do the work. That way, if the demand for that work drops, one can simply stop using the services of the external person or company rather than having to deal with firing employees whose services are no longer needed. 

 


 

 

 

Search this Site

TTG Home

Robert Home

Daynotes Home

Special Reports

Current Topics


Thursday, 26 August 1999

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]


I (finally) got off another chapter to O'Reilly yesterday, this one about Optical Storage. Now to start another one. Fortunately, I only have one more "hard" chapter to do, followed by a bunch of easy ones. I'm starting to see the train at the end of the tunnel.

And here's an unusual problem with rechargeable batteries. The ones in my Olympus D-400 Zoom are lasting too long. Barbara and I are heading out this morning for a trip to Asheville, where she has a one-day conference. Obviously, I want to take the digital camera, but it's *still* running on its original charge, this after taking about 150 pictures. I walked around the house yesterday shooting lots of flash pictures and using the LCD in an attempt to deplete the charge. No joy. 

I'm out of time for running down the battery before we leave, at least if I want to be able to recharge for the 100 minutes or so this charger requires. I think what I may do is just charge the damn things even though they're not completely run down. These NiMH batteries (yes, I know they're really "cells", but that sounds stupid) are supposedly less subject to memory effect than are NiCd batteries, so perhaps I won't hurt them. Of course, I also carry around a set of alkaline batteries for emergencies, so perhaps now is the time to give them a try. I've heard horror stories about the short life of alkalines in a digital camera, but haven't had the chance to see for myself yet.

We'll be up in Asheville all day today, staying at a bed and breakfast overnight, and coming back late tomorrow afternoon or evening. My brother is coming over from Raleigh to mom-sit, and we have various other people lined up in case there's any kind of emergency. Mom will have more people around here while we're gone than she ever does while we're at home. 

Hmm. The thought just struck me. I wonder if our B&B has a web site. If not, I can offer to set one up for them (at my usual outrageous billing rate, of course) and take some photographs of the place while I'm there. Speaking of outrageous rates, this place charges about $100 a night. When we used to stay at B&Bs ten and fifteen years ago, they were about a third the price of a decent hotel. No longer, apparently.

* * * * *

This from H [hstuck@excite.com]: 

IE 5.0 always used to ask me for the target directory for FTP downloads. However for the last month or so it has not. And since what it seems to be defaulting to is not close to the download directory I always have used, one needs to run a search program to find where it put the downloaded files. (Which are usually in a temp directory, that as I understand it, would disappear if I closed IE.)

Ugh. Be nice to get it back to always asking. I won't dispute that maybe I did something to trigger this behaviour, but I don't know what it was and don't know how to unchange this new behaviour.

Conclusion: I'm in agreement with your comments.

Sorry to hear about your problems, but I'm glad I'm not alone. I knew I wasn't, really, but from reading reviews of these products one wonders what the reviewers are thinking about. I have encountered many common bugs with IE, Navigator, FP2K, etc., but reviewers never mention them. Surely they must have encountered them if they used the product at all. Perhaps they're paid not to mention problems. I don't know.

* * * * *

This from Jan Swijsen [qjsw@oce.nl]: 

Well, I don't know enough about the government and business environment in Belgium to comment,....

I don't know the rules and situation very good either but I know about some examples and cases. It is much more difficult to fire an employee here in Belgium than it is over there. And when you fire him/her you have to pay a lot. The reason behind these laws is not bad, it prevents employers to dump and replace employees at whim. Especially for low schooled employees it provides some protection.

The principle is sound but, as so often with laws, the execution is FUBAR.

If they could hire and fire at will, ...

The problem is that the longer an employee is in a company the more holidays he gets and the higher the pay becomes. Typically we get a raise every year, the height of it is to be negotiated. We get an extra holiday every 5 years (I work 9 years for my company now so I have 22+1 holidays per year). If firing is easy and cheap everybody would be fired and rehired after 4 years. 

We need laws to protect employees just like we need laws to protect customers from fraudulent sales practices. It is however difficult to get effective protection without hindrance to normal correct business. 

If everybody would be nice and honest we wouldn't need any laws.

Well, I'm not sure where to even begin rebutting your remarks. Obviously, what we have here is the traditional European statist Keynesian viewpoint versus the traditional American free-market Austrian School viewpoint. First, I take strong exception to your statement that "the principle is sound". It isn't. The principle is the problem. Your argument reminds me of that old chestnut that says that the theory of communism is good and admirable, but that it has been poorly executed in practice. Crap. The very idea of communism is evil, and its defects when implemented flow inevitably from its defects in theory.

All that the European model does is establish a protected class--those who are already employed. It does nothing to "protect workers" and in fact damages their interests. People in this country who support the Minimum Wage laws use a similarly specious argument when they say that Minimum Wage guarantees that people will be able to work for a living wage. Crap again. All the Minimum Wage laws do is make it illegal to employ anyone at less than the minimum wage. Someone who is not worth that much to an employer simply can't find a job. That worker probably doesn't feel very protected.

By making it difficult or impossible for an employer to fire an unneeded worker, all that you accomplish is to ensure that that employer will hire as few workers as possible. The very fact that the force of law is required to sustain this situation proves that it is unnatural. You seem to assume that workers should automatically be given higher pay, longer vacations, and other benefits that are costly to employers simply because those employees have been with that employer for a certain time. Why? Is an assembly-line worker with one year's experience more valuable to the employer than one with one week's experience? Probably. Is a worker with five or ten years' experience more valuable than one with one year's experience? Probably not. Why, then, should the employer be forced to pay more for the long-term employee?

All this kind of practice ensures is that a company (or a country) becomes uncompetitive because it puts the cost of its labor at a higher level than the market will bear.

 


 

 

 

Search this Site

TTG Home

Robert Home

Daynotes Home

Special Reports

Current Topics


Friday, 27 August 1999

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]


Out of town. Nothing new until Saturday.

 

 


 

 

 

Search this Site

TTG Home

Robert Home

Daynotes Home

Special Reports

Current Topics


Saturday, 28 August 1999

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]


My wife says I'm weird, and that this goes to prove it. For several months, I've been plagued by a persistent case of Athlete's Foot. Why, I can't imagine, since I avoid exercise at all costs. I tried everything. Tolnaftate, miconazole, and then other assorted gunk with various active ingredients, including one that had been available only by prescription until recently. I used them as directed. I used them longer than directed. Nothing worked. I finally tried combining them, but that didn't work either. 

My wife said I should go to the doctor, but I haven't been to a doctor for at least 25 years and didn't want to break my string of successfully avoiding them. (She did convince me to get my blood pressure checked at a booth set up by some EMTs in a shopping mall ten years ago or so. My pulse was 61 and my blood pressure was 90 over 60--that's what playing tennis in my youth did for me, I guess).

At any rate, none of this gunk was working and I sure didn't want to visit a doctor. Nasty places, doctors' offices. There are sharp, pointy things there. So I decided to roll out the tactical nukes. My reasoning went (a) this is a topical problem, for which I've been applying topical medicine, (b) there certainly must be a more effective topical anti-fungal agent readily available, (c) potassium permanganate or sodium hypochlorite, aka Clorox, are strong oxidizing agents that will kill anything on contact. So, I announced to Barbara that I planned to soak my feet in laundry bleach. She said I was nuts.

Thinking that perhaps she was right, I decided to do a little Internet research. Sure enough, both sodium hypochlorite and potassium permanganate showed up in various resources as treatments for various fungal infections, including Athlete's Foot. But one intriguing thing that I hadn't thought of also showed up. One source suggested using SoftScrub with Clorox, a liquid abrasive cleanser that contains a mild abrasive and laundry bleach. So I tried it. One application helped a lot. By the second application, the problem had just about disappeared. I am not a doctor, but I think I'll try this method first next time...

* * * * *

This from Jan Swijsen [qjsw@oce.nl], with my comments embedded:

Well, I'm not sure where to even begin rebutting your remarks.

At the top usually is best. 

statement that "the principle is sound". It isn't. The principle is the problem.

The principle I meant has nothing to do with communism. I meant that there is a need for protection from fraudulent practices. Ex payment is at the end of the month (custom here) you get fired on the 29th and don't get your pay for the month because it would have been paid the 30th but you no longer work then. Presto your boss got about one month free labour. You got no pay. 

I think protection is needed, maybe you don't. 

I wasn't accusing you of being a communist, merely saying that your assumption reminded me of that old thing about communism in theory being good but communism in practice being bad. Communism is bad in both theory and practice, and I wasn't conceding your assumption.

As far as the example you note, we were obviously talking at cross purposes. If an employer makes use of an employee's services and then refuses to pay for those services, that's fraud. An employer couldn't get away with that in this country, nor would I want an employer to be able to get away with that.

By making it difficult or impossible for an employer to fire an unneeded worker, all that you accomplish is to ensure that that employer will hire as few workers as possible. The very fact that the force of law is required to sustain this situation proves that it is unnatural.

Every employer is hiring as few workers as possible, by economic law, no government law required here. Unneeded workers are few indeed. If a short term extra employee is needed one can be contracted using a limited contract. What is made (overly) difficult it to fire a worker with a normal contract. Firing at rem would be a breach of contract and the safeguards built into most contracts are (too) severe. 

Most contracts were negotiated long ago and it seems to be difficult to get rid of rules afterwards. Adding new stipulations and restrictions is easy. In theory the contracts should reflect a fair and economically correct method. In practice they lag about 5 years behind the economic situation. And they apply on a sector level. If most companies in a sector (ex textiles) are doing well and one has a problem then that problem is made more severe. 

Of course any employer attempts to avoid hiring unnecessary employees, but that wasn't my point. What I was saying is that because it is so hard to fire an employee, employers hire fewer employees than they actually need, and make up the difference by contracting outside. In the absence of government regulation, employers would hire the employees they needed to do the work that needed to be done. With government regulation, employers use contract workers and temps instead, even when it would otherwise be more efficient and more effective to make direct hires.

You seem to assume that workers should automatically be given higher pay, longer vacations, and other benefits that are costly to employers simply because those employees have been with that employer for a certain time.

I don't assume this, personally I don't take up all the holidays that I an entitled to (but I am not typical in this). A few years back, we even had laws that forbade pay rises!! (to get the inflation down prior to being acceptable for entry in the Euro system). Sounds incredible but true. The only way to get a raise was getting another job or function (I changed from Programmer-analyst to Analyst-programmer ;~) ) 

It is a common 'rule' for office employees and the raise is not assured, it has to be negotiated individually. In offices it is not uncommon to have two people doing the same job (equally good) but getting different payment. I don't know if this is also the practice for assembly line workers. On assembly lines typically everybody doing the same job gets the same pay (after a few weeks introduction). 

These 'rules' are not laws mandated by the government but rather negotiated between the workers unions and the employers 'unions'. These negotiations occur on sector level, so, for example all car manufacturers have to apply the same rules (mind you, not the same wages). That they have to comply with the rules is government law, the rules themselves are not 'government laws' in the sense that they are not dictated by someone in Brussels (Washington for you). They are the result of negotiations between employees (represented by unions) and employers (also grouped). 

I think that the sector wide nature is what makes the most difference between the American view (not the actual American situation) and the 'European' view (not the act...). 

I like communism. 

In other countries. 

In this country, two major methods are used to determine pay raises and level of benefits. 

The first, "seniority", is used primarily in jobs where people are interchangeable cogs--e.g. production lines, manual labor, etc. That is, places where individual differences in employee performance are either minor or too difficult to track. But it is here that pay raises make no sense (excepting those that adjust for inflation). A guy who's been bolting on fenders for five years is no more efficient or effective at his job than one who has been bolting on fenders for a year. In jobs like this, everyone should be making the same amount of money regardless of how long they've been doing the job.

The second, "merit", is used for jobs where individual performance both varies significantly and is relatively easier to evaluate and reward directly. There are pitfalls with this method, too, not least the variability of the people doing the evaluations, but when implemented properly this method at least takes into account the value of the person being evaluated to the organization.

The difficulty with merit raises is that, although they can function well in small organizations, it is very difficult to implement them properly (and fairly) in larger organizations. The result, in this country at least, has been the introduction of something called "pay ranges", where a given type of job is assigned to a pay range and managers award merit raises within that range to individual employees. The problem with that method, of course, is that the pay ranges assigned to different types of work are often nearly arbitrary. For example, because the personnel department usually sets those ranges, one often finds that personnel clerks (who should be at a very low pay range based on the small skills needed for the position) are often equated with skilled positions like computer network analysts. The result is that you have payroll clerks hanging onto their jobs for dear life, and computer analysts looking for jobs elsewhere.

* * * * *

This from Daniel Seto [dkseto@email.com]: 

I agree with most of what you've said regarding hiring and firing of employees in the private sector (we can leave to another day a discussion of the public sector). I absolutely believe that a private business must be able to hire and fire, at will, and not have to justify it to anyone. Least of all people who are not even customers.

I further believe in merit pay and not raises that are solely tied to longevity. However, I needed clarification on the following:

"You seem to assume that workers should automatically be given higher pay, longer vacations, and other benefits that are costly to employers simply because those employees have been with that employer for a certain time. Why? Is an assembly-line worker with one year's experience more valuable to the employer than one with one week's experience? Probably. Is a worker with five or ten years' experience more valuable than one with one year's experience? Probably not. Why, then, should the employer be forced to pay more for the long-term employee?"

Bearing in mind that I don't think automatic pay increases are a good thing overall, your example seems to equate experience with longevity and that there shouldn't be any pay raises at all, which some advocate, based on experience (versus longevity).

Is anyone with one year of experience more valuable than one with one week? Probably. Is anyone with five or ten years of experience more valuable than someone with one? I certainly hope so! And if not, why is that person still employed there? And whose "fault" is it if they are?

So, if we agree that there shouldn't be automatic increases based longevity, does that mean longevity is the same as experience (I don't think so)? And if it isn't the same, does that mean experience should be excluded in toto as one criterion for a raise?

Aloha,

I think we have to differentiate between different types of jobs. I was going to use the words "skilled" and "unskilled", but those have taken on meanings that no longer reflect reality. For example, the guy who bolts on fenders all day every day is considered a "skilled" laborer, but that's not really the case. Anyone who does a job that is a candidate for replacement by a robot is not doing a skilled job. 

The guy who's been bolting on fenders for five years is worth no more to the company than someone who's been doing it for three months. After a certain (short) period of doing that job, one doesn't do it any better or any faster, and one shouldn't be paid any more simply because one has been doing it for longer than the guy in the next bay. As far as experience, the guy with ten years on the job doesn't have ten years of experience. He has three months of experience forty times.

Skilled jobs, on the other hand, should be rewarded according to performance (not experience), and are in a free market.

* * * * *

This from Joshua D. Boyd [jdboyd@cs.millersv.edu]: 

I've recently been hearing reports that the ATI All-In_Wonder 128 comes with a digital VCR program, and that it uses real mpeg compression. It also has TV out I believe. Thought you might be interested. 

Anyway, I had a question I wondered if you could help me with. I know that you aren't really into AMD chips, but a friend recently pointed out that some of the K6-2 chips are really cheap (333mhz for $35). I was wondering how one would go about figuring out if their mother board would support such a chip? Like you, I've never really been into AMD, but if I can upgrade my Pentium 200 to a K6-2 with 3D Now for only thirty odd dollars, that would be a worth while upgrade. Especially since the software I use that needs high performance is mostly custom written, so I could potentially rewrite it to use 3D Now, which is really far better than MMX (MMX is fairly close to useless for 3D graphics because it only works with integers and has that awful hit when switch between FP and MMX modes).

The fastest CPU that my mother board mentions supporting is a Pentium 200. The mother board is over 2 years old. I know that it wouldn't be able to fully support an AMD chip meant for a 100 mhz super 7 board, but how what I determine what other AMDs it would support. It mentions support for AMD K6s up to 166, which were the fastest available at the time the board was made. Is the only issue for faster CPUs mother board voltage, or is there more?

I know that you don't really like AMDs, but then, you seem to be comparing K6-2s and 3s to P6 core chips, whereas I'm comparing them to P5 core chips, and as a cheap upgrade, not a whole new system. If you don't have the answers to my question, do you have any suggestions on where to find them?

--

Joshua Boyd
http://catpro.dragonfire.net/joshua

Your current motherboard may or may not support the AMD K6-2. The K6-2/333 is probably not a good choice because it uses the rather odd 95 MHz bus speed, which your motherboard probably doesn't support. The best thing to do is contact the motherboard manufacturer to determine whether that board supports the K6-2, and if so at which speeds. There are several issues:

1. Chipset. Determine which chipset the motherboard uses. Early chipsets simply don't provide the essentials needed to run the AMD K6-2. If you have an Intel 430TX or a late-model Via, ALi, or SiS chipset, you should be okay.

2. Bus speed. Most of the K6-2 CPUs run at 100 MHz, so your motherboard must be configurable to use that speed.

3. Voltages. K6-2 CPUs are split-voltage processors. They require 3.3 V I/O and 2.2 V core. Your motherboard must be able to provide both of those voltages.

4. Multipliers. If your current motherboard specifies a 200 MHz Pentium as the fastest processor it supports, that may mean that it may be jumperable to a maximum of 66 MHz with a 3X multiplier. But that may also mean that the 200 MHz Pentium was the fastest CPU available at the time, and you may in fact have faster multipliers sitting unused on the motherboard.

5. BIOS. If you have an older BIOS, support for the K6-2 may be non-existent or problematic. If you have a Flash BIOS, check to see if an update is available.

 


 

 

 

Search this Site

TTG Home

Robert Home

Daynotes Home

Special Reports

Current Topics


Sunday, 29 August 1999

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]


Slow day today. Barbara just left on a trip to meet another woman at a Dairy Queen in Virginia, where she'll deliver Bonnie, the Border Collie we've been keeping for the last couple of days.

* * * * *

I'm busy downloading WinAmp 2.50c, the popular MP3 player, which is now freeware. I seem to remember downloading an eval copy of an earlier version some time ago, when it was still a shareware product, and finding that it wouldn't play MP3s recorded at 256 Kb/s. If that's the case, it'll be useless for me, but I figured it was worth checking. The ripper that comes with Plextor Manager 2000 is first-rate, and Tord Jansson's BladeEnc MP3 encoder is fast, free, and produces very high quality MP3s at high bit rates, which is what I'm interested in. I won't do any production ripping until I have that gigantic Seagate 50 GB drive running in a server, but I do plan to play around some with MP3s first.

* * * * *

And here are a couple of letters offering help on the matter of upgrading an older Pentium system to use an AMD K6-2.

* * * * *

This from bdenman [bdenman@FTC-I.NET]: 

Saw your comments in Bob's column today and decided to pass along my experience as an AMD user (currently K6-2/400, formerly 300).

- The K6/2 series runs either at 100MHz or 66MHz (assume the 95MHz version can run at 95 or 66). Some K6 were actually made only for 66MHz but they are not common. By changing my boards multiplier and FSB speed, I have run the 300 at 100 or 4.5 x 66. This is legal too; no shenanigans. You can run chips slower so the 333 chip should run fine at 4.5 x 66 without losing much.

- Bob is right on the money on the dual voltage thing. Don't try to use a K6 in a Classic Pentium board. Repeat....Don't try to use a K6 in a Classic Pentium board. You will fry the chip.

- Intel went to a dual voltage chip too in the P-200 MMX and above. A two year old mother board sold with classic Pentium chip probably will not support dual voltages. But, check your motherboard manual and see if it will support a MMX chip. Also see if it has a multiplier of 4.5 (doubtful). A current BIOS would be helpful but is not necessarily mandatory (caveats apply; Murphy's law applies; you might be risking $35). As long as the chips runs okay what the BIOS reports as chip type is not important. 

- Recommend you buy new motherboard that will use the inexpensive K6-2 with your standard DIMM unless you want to buy PC-100 DIMMS. My ATX board does that; cost me $69 plus shipping. The AT version was $59. Shipping was reasonable (see recommendation below).

- Regardless of Bob's recent opinions and Anand's recent silly statement regarding no stable SS7 motherboards till now, IMHO TMC makes a decent stable board for the K6. And has since last year. Proof is in using one for almost a year.

Note: I am on third one: I gave my first m/b, AT version, to a son in Alaska along with my K6-2/300. (he needed stable system; Alaska is long way for tech support from South Carolina). He is using with 32MB EDO SIMMS at 66MHz. I replaced it with one just like it but with a 400MHz chip. This motherboard I passed along to my son-in-law for him to run with his K6-2/300 (technically a 66MHz variety btw but it runs okay at 100MHz). I moved to an ATX case which prompted the third board. I now run run a K6-2/400.

My source for these boards btw was Microbarn; a mail order outfit out of North Carolina. I have had good luck with them personally. The AT version of board called AI5VG+, for ATX: the TI5VG+. I am sure there are other boards out there that will work just as well. 

Holler if you have further questions. Not sure where you are; but is you would like to call that would be fine. Will be home this evening. 

Bruce Denman
bdenman@ftc-i.net
(803) 481-5174 (stay on line if machine picks up and say who you are/why; we have to screen calls unfortunately)
bdenman@ftc-i.net
http://web.infoave.net/~bdenman

* * * * *

This from Dave Farquhar [farquhar@access2k1.net]: 

By all means check out [this link and this link (URLs converted to links to prevent horizontal scrolling) ed.] before doing anything with K6-2s. The latter link deals with K6-IIIs, but the principles are pretty much the same. If a board will work with a III, it should work with a 2.

Bus speeds aren't much of anything to worry about--just take a bus speed you do have, then find a multiplier that takes you close to the CPU's speed. You could clock a 333 at 66 MHz x 5 (330 MHz) or 75 MHz x 4.5 (337.5 MHz). Remember that if you stay within a 2.5 percent margin of error, overclocking isn't a big deal--the chips are designed to withstand those kinds of variances. But even if the best you can manage is 300 MHz with your board, it's not the end of the world when these things are selling for 35 bucks.

You might also take a look at IDT WinChip 2 (if you can find them) or Cyrix CPUs. It's a lot easier to get one of those to work, but they both suffer from weak floating-point performance. The WinChip 2 does have 3DNow! support, which helps. The WinChip 2 uses similar voltage settings to a Pentium MMX (it is a split-voltage chip, unlike the original WinChip), as do Cyrix chips.

Supposedly the K6-2 works fine at 2.5 volts (it's rated for 2.2), but I haven't been gutsy enough to try my K6-2 at that voltage. I may not have paid much for that chip but I'd really rather not blow it up because I've come to like it. I have an old Abit IT5H motherboard that I've been meaning to drop a Cyrix chip into in order to extend its useful life a bit. (I don't do much floating-point intensive stuff.)

I hope this helps.

Dave Farquhar

 

 

Copyright 1998, 1999, 2000, 2001, 2002, 2003, 2004 by Robert Bruce Thompson. All Rights Reserved.