jp.jpg (13389 bytes)

CHAOS MANOR MAIL

A SELECTION

December 20 - 26, 1999

REFRESH/RELOAD EARLY AND OFTEN!

read book now

HOME

VIEW

MAIL

Columns

BOOK Reviews

emailblimp.gif (23130 bytes)mailto:jerryp@jerrypournelle.com

CLICK ON THE BLIMP TO SEND MAIL TO ME

The current page will always have the name currentmail.html and may be bookmarked. For previous weeks, go to the MAIL HOME PAGE.

If you are not paying for this place, click here...

IF YOU SEND MAIL it may be published; if you want it private SAY SO AT THE TOP of the mail. I try to respect confidences, but there is only me, and this is Chaos Manor. If you want a mail address other than the one from which you sent the mail to appear, PUT THAT AT THE END OF THE LETTER as a signature.

I try to answer mail, but mostly I can't get to all of it. I read it all, although not always the instant it comes in. I do have books to write too...  I am reminded of H. P. Lovecraft who slowly starved to death while answering fan mail. 

Day-by-day...
Monday -- Tuesday -- Wednesday -- Thursday -- Friday -- Saturday -- Sunday

Search: type in string and press return.

 


Boiler Plate:

If you want to PAY FOR THIS there are problems, but I keep the latest HERE. I'm trying. MY THANKS to all of you who sent money.  Some of you went to a lot of trouble to send money from overseas. Thank you! There are also some new payment methods. I am preparing a special (electronic) mailing to all those who paid: there will be a couple of these. I am also toying with the notion of a subscriber section of the page. LET ME KNOW your thoughts.
.

If you subscribed:

atom.gif (1053 bytes) CLICK HERE for a Special Request.

If you didn't and haven't, why not?

If this seems a lot about paying think of it as the Subscription Drive Nag. You'll see more.

Highlights this week:

 

line6.gif (917 bytes)

This week:

Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday

TOP

Monday December 20, 1999

Opening a new alt.mail topic in infantry and branch of service.

Be sure to see Sunday 19th Mail on many topics.

So we'll go no more a-roving...

According to Alan MacRobert of _Sky &; Telescope_ (in an article in the 12/20 Boston Globe), the mail going around about the exceptionally bright moon is (mostly) an Internet myth:

o Wednesday's moon will be about 19% brighter than usual: detectable to the naked eye only if you had both moons side-by-side.

 o There were brighter moons several times this century, e.g. January 15, 1930, and January 4, 1912. o The story originated in the Old Farmer's Almanac, but was unnoticed until somebody put it out on the net.

Fascinating. I presume the astronomical events are true, but the perceptions won't be as large. Sigh. Well, the heart outwears the breast...


It appears that competition is finally starting to lower the cost of registering domain names. Has anyone registered a domain with any of the alternative registrars? I see that CORE (http://www.joker.com) is registering .com domain names for $37 for the first two years and $19 annually thereafter.

-- Robert Bruce Thompson thompson@ttgnet.com http://www.ttgnet.com

I have sent my renewals in to the Internic in the usual manner, although their arrogant manner annoys me. It's not enough money to worry about, but perhaps with competition they may think of giving some "service" for collecting the money. They have the most arrogantly ignorant on-line registration "service" I can imagine, with everything having to be EXACTLY their way. Time they got some competition.


Jerry,

I have a Yamaha SCSI CRW4416 CD writer on my Linux box at work (one of the perks of being an MIS guy), and I've gotten lots of use out of it over the past couple of months.

The CD writer has worked out well for me - especially when I'm making new Red Hat CDs. I bought a copy of 6.1 solely so I could have access to Red Hat's priority FTP update site, and I grab the latest RPMs as soon as they become available. Then I run a little script which replaces old RPMs with new ones (I have the full install tree on my hard drive), then creates a new ISO CD image which I can burn onto a CD-RW whenever I need to install Linux outside the company (onsite installs are always done via the network).

I do this sort of thing via the command line, since it's the fastest way (for me) - but packages like xcdroast let you do all this via a GUI.

I don't know about IDE CD writing under Linux, but I'll probably find out, since someone donated such a device to my church, and it will go into the Linux server there (that's the only computer with the horsepower to do that - everything else is old, donated equipment, and the fastest desktop is a 60 MHz Pentium).

BTW, I think Diamond Multimedia is getting a clue about Linux. Why? Well, my employer recently bought a Dell Precision 410 system for one of our engineering people, and I got to set it up - including Linux, at her request. This computer has a Diamond FireGL1 video card (which retails for about $1000), and drivers are available only for NT AND Linux (beta, but it seems to work pretty nicely at 1600 by 1200 on a 21 inch monitor) - NOT Win9x.

Calvin Dodge

p.s. I'm enjoying the military discussions on your web site

I have been enough under the weather that I won't be trying this for a few days. Thanks.

I have had cdr support in linux for a long time. At work I use a combination of a dvd drive and a 4x ide writer. At home the combination is a 48x ide cdrom with an ide 2x hp writer. In both cases it works great. I will admit that getting this combination to work is not the easiest.

If I remember right( last time I set this up was 4 months ago) I installed the rpm file of a program called xcdroast and recompiled my kernel. It is all laid out in the cdr howto at http://www.cc.gatech.edu/linux/LDP/HOWTO/CD-Writing-HOWTO.html  . If you decide you want to get this to work, just email me some details about your setup like distribution, what device the cdrom is going to be, and kernel. I can then send you some detailed instructions.

CDR support in linux is better than in windows. I have never had a buffer underrun in linux. The programs are a little less friendly but more powerful.

Piotr CS System Support, University of Georgia. misztal@cs.uga.edu | 

 


SAMe: A Caution

Jerry,

There is one fact about taking supplemental SAMe that you need to be aware of. SAMe is broken down in the body to homocysteine. If the body has enough cofactors (these include folic acid, and B-6) to metabolize the homocysteine properly, than there is no problem. However elevated levels of homocysteine are much worse than elevated cholesterol in the damage to blood vessels and accelerated atherosclerosis that they produce. The best way to make sure your homocysteine is not elevated is to have a blood test to measure it (any elevation above normal is bad). A minimum supplement regimen to prevent elevated homocysteine would include 25 mg of most of the B vitamins (it is important to not just take one or two but all the B vitamins) and 400 micrograms of folic acid. Better would be 50 mg of most of the B Vitamins and 800-1600 mg of folic acid. Some people might have elevated levels of homocysteine even with this regimen, so the safest bet is get the blood test.

Public Disclaimer: Please consult with your physician before starting/changing any vitamin regimen, as vitamins can interact with medications you may be taking.

Scott La Pidus, M.D. AScottLP@aol.com

Thank you. I always take a B-100 with 800 folic acid as well as my usual immune protectors with SAMe. And I don't recommend that anyone try what I do; I merely report. I do seem to have recovered from this local flu about as well as anyone else, and given my age, that's not bad thing, especially since I got a few thousand words written while under it, so something must be going right...   Thanks again


In last weeks mail we looked at digital movies. A continuing discussion:

Digital Movies: What Ebert overlooked.

Dear Dr. Pournelle,

I, too, read Mr. Ebert's condemnation of digital projection. While I have not yet had the opportunity to see a digitally projected film myself, I can see one glaring omission in Mr. Ebert's consideration of the topic.

Mr. Ebert's objections fail to consider the abysmal quality of today's multiplex projectionists.

In many theaters in the U.S. and Canada, there are no longer any full-time projectionists. In the old days, a knowledgeable projectionist would run two projectors, swapping reels and carefully maintaining the machines. If there were any problems, the projectionist could fix them.

Today, the projectors are automated, and many theatre chains have done away with the projectionist. Instead, there's a "film technician" who comes in to set up new films, or move films to a different theater. This technician may need to cover several theaters in one city. Maintenance becomes an on-call issue. The theater owners figure that automation means all they need is a manager (or even a popcorn pusher) to flip the switch at the right time.

As a result, the quality of today's movie presentations is abysmal and getting worse. Hairs and dirt on the film? Commonplace, in my experience, even on brand new films. I've seen plenty of films that are terribly scratched up within a week of release. I've seen films whose soundtracks were mangled by poor threading. I've stopped complaining about the "digital sound" trailers that are so old and so worn that the sound drops out entirely halfway through, because the soundtrack is almost gone.

Yes, we now have digital sound for film -- but it's a hack, and not a very good one. The most reliable digital film sound is probably DTS -- but that requires CD-ROMs to be shipped with the film, just like the old Victrola days with wax records. Often, the CDs are missing by the second run. Dolby Digital and SDDS have to make due with the leftover bits of film near the sprocket holes, where they are very subsceptible to damage if the film isn't properly handled. Between the extra stress that the automated "platter" system puts on the film, the lack of daily skilled maintenance, and the hurried film tech, those on-the-film soundtracks often have a lifetime measured in weeks.

It gets worse as the multiplex gets greedier. Sometimes, if a multiplex wants to run a popular movie in more than one theater, they will not get two prints of the film, but rather "chain" one print. That means taking the film off of the automated platter, running it through one projector, then through a second projector in the second room (possibly crossing a hallway in the process) and finally back to the platter. That's a lot of stress for a thin strip of acetate -- and lots of rollers to gouge the film.

Finally, multiplexes have started building bigger and bigger auditoriums, with stadium seating and room for huge opening day crowds. In more and more cases, the screen has become so large that it is no longer possible to get a sufficiently bright projected image. There's only so much light you can pump through a 35mm piece of film before the heat causes it to melt. That's one reason why 70mm film was invented, and why it remains popular in "special venues" with large screens: you can put more light through it without melting it, because of the larger surface area.

Mr. Ebert apparently discounted all of these real-world problems of film. Yes, in an ideal world, we would all be movie critics who get to see pristine copies of films lovingly tended by a union projectionist. However, most of us are at the mercy of the major cinema chains, and we have to live with dim, scarred, garbled presentations.

Digital cinema, even in its current embryonic state, addresses these problems. Digital projection, especially with DLP, is very bright. The micromirrors on a DLP chip can handle a lot more heat than 35mm film. Digital projection is not subject to scratches and torn sprockets and framing problems. Digital projection doesn't suffer if you "chain" the film. Digital projection has an integrated digital soundtrack that isn't inherently prone to failure.

I understand Mr. Ebert's dismay at film being replaced with something that has been called, at best, "as good as" film in terms of optimal image quality. However, I would gladly put up with "as good as film" if it meant that I could still see "as good as film" after a movie had been playing for six weeks, or in a second-run theater. With film, I know that I have to go the first weekend if I want to see the movie before it gets all beat up.

That, in itself, makes digital projection desirable and likely to eventually kill celluloid. It's also a sure bet that the resolution will improve to best 35mm film by the time that the projectors become widespread.

As for Mr. Ebert's arguments about bandwidth, I see no reason why films could not be shipped on DLT, 8mm AIT, or even 12" optical platters. Consider that a single-sided DVD can hold about 9Gb of data, and its predecessor (CD-ROM) held only 650Mb. One more generation of refinement in that technology ought to take care of the cinematic movie storage problem handily.

-- Rob Levandowski robl@macwhiz.com

I saw Tarzan with digital projection and I had no objections or qualms about the quality whatever. This was in Burbank. I did a short column on the coming effect of digital in general (bits is bits) for Intellectual Capital a few months ago, and I think it will have a great impact on the entertainment industry. You raise interesting points.


Continuing from last week:

> The real question on the JAVA Apps is, are Microsoft's portable > apps slower than anyone else's? And are other people's as fast > as the Microsoft non-portable? And finally, does anyone else make > "Windows Only" apps and if so are they as fast as Microsoft's > "Windows only"? You see where I am headed.

In the benchmarking which I did, Microsoft's Visual J ran portable apps as fast as anybody. The problem is, that it's too much bother to make the portable apps in Visual J.

I haven't seen anybody else producing non-portable code.

The situation, as I see it, goes something like this:

The main attraction of Java, to many users, is it's portability. In some cases that reflects a real need to operate on multiple systems. In many cases it's simply a desire for non-Microsoft operating systems to be made more viable. One of the biggest proponents of Java is Sun. Sun would like to sell many more desktop units, but finds it hard to compete as there are so many Windows-based apps in thge marketplace. We get a Catch-22 effect (RIP Mr Keller) where developers won't write Sun software as there aren't enough customers, but customers won't switch to Sun as there aren't enough apps. If Java becomes the language of choice for developers, then developers can sell software to Sun customers (good for the developers) and Sun can sell more hardware (good for Sun). The whole deal is good for everybody, except Microsoft who will lose sales.

Now Microsoft, who are seldom slow in analyzing the market, release their own version of Java which is just for Windows computers. This offers developers the advantages of developing in Java (which are considered by some to be quite significant), but without the poor performance which has dogged portable Java. There are many non-sinister reasons why this is a good move for Microsoft; and if it foils that nasty drift away from Windows-based computers then so much the better.

Sun, not unexpectedly, are suing Microsoft. They claim that what Microsoft is selling is a different product from their copyrighted "Java" product - so Microsoft ought not use the name.

I guess that by the time the courts resolve that, the battle will probably have been decided in the marketplace.

Depending on your personal philosophy, you can read this as "another example of manipulative behavior" or "some pretty clever marketing".

I'd personally like to see more vigorous competition in the OS market, so I'd like to see portable Java succeed - if only to spur Microsoft into working harder on the Desktop market. (They've recently neglected that pretty badly as they concentrate on the Server market). However I'll avoid joining the "Microsoft==Anti-Christ" chorus for the foreseeable future.

 Michael Smith emmenjay@zip.com.au Emmenjay Consulting Pty Ltd http://www.zip.com.au/~emmenjay/

Just when you've made it foolproof, along comes a better class of fool.

I agree completely that the marketplace will decide the issue. Thanks!


What's the brand and URL for the SCSI cables you feel are head and shoulders above the rest? I used to have it noted somewhere, but now can't find it...

Thanks.

Tony Rairden First Quality Musical Supplies www.fqms.com

Granite Digital


 

 

TOP

 

This week:

Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday

TOP

Tuesday, December 21, 1999

I have a friend who programs in Java. He's a Sun Solaris fan for almost everything, but programs Java for his clients on an NT box using Visual J++. I asked him why and his response was quite intriguing. He says it takes at least 50 times longer to program in Java, so he makes 50 times as much money. An application he could knock up in a day or so using Visual Basic will take many weeks in Java. He says he can get away with this because of Sun's hype about Java and anti MS sentiment.

I asked why use the MS product when he could be using something from Sun. Was he programming for Windows only and looking for the extra speed from the Win API hooks? No, he said that Visual J++ was a much more efficient environment to program in and that MS making it harder to program "run anywhere" Java apps was total BS. He said that VJ++ was so much better than other Java development tools, that it is much quicker to develop in.

Quite amusing, I thought.

Jonathan Sturm

Provocative also. I have always known that the real purpose of C is to employ C programmers; since the number of times any C code has been transferred successfully from one programming team to another is very nearly nil, if you do your project in C you will have a job for life or at least the life of the project. Is Java that way also? Well, we ought to get some mail on this...


I think he's right. I'm a software engineer. I work at Rockwell- Collins In-Flight Passenger Entertainment Systems. (Opinions here are my own, of course.) People always prefer better quality.

It's pretty easy to calculate the bit rate for a perfect digital movie media. First, film is better than the human eye, or the cinematographer. Resolution of film is limited by the proiector's frame-alignment error. Cinematography is not perfect, but I know nothing about it, so let's give the cinematographers perfection based on human factors.

The minimum angular resolution of the human eye is ~0.0136 degrees. Make the movie 90 degrees wide, and 60 high for classic perspective, not Cinemascope. Cinemascope is a compromise to make the medium more immersive wihtout a major increase in auditorium costs. With a fast monitor some people can see a flicker out to about 40 Hz. Nyquist's theorem doesn't exactly apply, but I took 80 Hz because I need a 70 Hz noninterlaced monitor to avoid headaches.

With 90 minutes the classic compromise between plot and bathroom, the gross uncompressed space for a perfect digital movie is 6.3e12 bytes. A combination of cosine compression, frame deltas, clever color transforms etc. can probably shrink this by an order of magnitide, to 630GB.

If DVD stores 5.5Gb, and Moore's law applies, (doubling capacity every 18 months) Maxivision 48 has about 10 years after DVD's market penetration, before utlimate digital media remove their market. In particular, Maxivision 48 pays -now- to use , and it will let the theatre owners leverage their existing equipment (including those lovely automated reels) after digital obsolesces it. Until it depreciates (probably another 5 years, or 10 if the buildings obsolesce).

Nanotechnic mechanically-sensed drives would have this capacity.

So, the Maxivision 48 has a product lifetime of 15-20 years. Which is perfectly reasonable. It's also a clear hack on existing technology, so existing users and providers will be comfortable with it.

Of course clever hacks can reduce the above size. However we were designing a premium lossless digital system.

Anyway, best wishes,

rgvande5@collins.rockwell.com

A reasonable analysis. One major advantage to digital distribution is reaction time: you can have a movie on far more screens if it takes off. With film you have to make and distribute the actual prints; with bits, it's a lot easier. Spinning metal disk drives lasted far longer than I expected, but there has to come a point when holographics or other silicon storage systems are cheaper and simpler. 

You may well be right that Maxivision 48 will give analog projection a few more years. I'll see if I can get my film experts interested in this discussion. My only observation was that Tarzan as digital projection was plenty good enough for me....


Jerry,

First off, let me say that I've been reading your books as well as Larry Niven's for years, and that it was a surprise to me when I found out that you wrote for Byte. (This was probably ~1996 or so. As an aside, Larry's books take up more room in my bookcase -- about 3 feet when I last checked -- than any other author...)

Moving on to CD burning on Linux, I've been doing it for about 18 months, under kernels 2.0.x and 2.2.x. Since I first burned CDs on what is now my main server (an overclocked AMD 5x86-150 box), I didn't have enough CPU power to burn on a GUI, so I used mkisofs to create the ISO image (a command line program, included with most distributions, which will make an ISO image from one or more directory trees) and cdrecord to burn it. (The CD-R I used the first time was an external SCSI Yamaha CDR-400t, which I borrowed for a few days.) This worked fine for data, but it would only create track-at-once audio discs. After I gave the CD-R drive back to the guy I borrowed it from, I didn't do it again until February of this year, when I bought a used Yamaha CDE-102 (external version of the CDR-102).

This new (to me) writer refused to work with the VLB Future Domain 18c30 chip on the SCSI card in my 5x86-150 server, so I attached it to a VLB Adaptec 2840 card in a 486DX2-66 box I had lying around. It worked fine there, with the same programs I'd used before (mkisofs to make the ISO image, cdrecord to burn it) -- I even tested writing a disc with the ISO image being NFS-mounted over a token ring network (yes, I'll find uses for old stuff others are getting rid of! :) ) and it worked fine.

A while later, I had a need to burn a disc-at-once audio CD. Since cdrecord wouldn't do disc-at-once, I tried various DOS programs, none of which worked. I eventually found cdrdao, a free program which *will* write disc-at-once audio CDs, and also allows me to change all the flags (2-channel/4-channel, infinitely copiable/copy once only <for whacked equipment which obeys SCMS>, and preemphasis. I tried it, and it worked very well. (The author pointed me to a new development version at the time, which had just recently added support for my CD-R drive.) Since then, I've moved on to using my current box (a 500MHz dual Celeron box using an Abit BP6 MB) as my main workstation, which finally has enough power to run X at the same time as it's burning a CD. (It still takes a fair amount of CPU time, though, as I'm using an old PIO-mode SCSI card I bought for $5 until I can get around to getting a real card. Still, it works very well for $5...)

If you have a SCSI CD-R drive, you just need to be sure that your kernel has SCSI Generic support compiled in or as a module (the module is called sg, I believe) and that you have read/write access to the appropriate device node in /dev. (The generic devices are /dev/sg<x>, with the numbers being assigned in the order the devices are listed in /proc/scsi/scsi.) If you have an IDE CD-R drive, ide-scsi emulation needs to be compiled in, and IDE CD-ROM support should not be. (SCSI CD-ROM support is needed, because with this module loaded, anything on the IDE bus which uses ATAPI appears to be using SCSI.)

If you're using CD-RW, there is no support, yet, for writing in packet mode. Some preliminary UDF filesystem support is available which may allow you to read a packet-mode disc, but as I don't have RW capability, I haven't tried it. If you want, you can use CD-RW discs for burning regular ISO-format discs, erasing them with cdrecord. (It includes a blank option which will either blank just the TOC, or the entire disc.)

If you have any other specific questions, feel free to ask away. I'll include links to some Linux CD-R stuff as well...

David Carter

Some links:

cdrecord: http://www.fokus.gmd.de/research/cc/glone/employees/joerg.schilling/private/cdrecord.html

cdrdao: http://www.ping.de/sites/daneb/cdrdao.html

cdparanoia (an excellent digital audio extraction tool): http://www.xiph.org/paranoia/index.html

A fairly good page of UNIX CD burning links: http://www.fokus.gmd.de/research/cc/glone/employees/joerg.schilling/private/cdb.html

-- David Carter ** dcarter@sigfs.org ** dcarter@visi.com PGP Key 581CBE61: E07EE199C767C752 8A8B1A9F015BF2EA Key available by finger or www.keyserver.net

My thanks, but I fear your letter causes me to think I need to re-open the Linux pages instead of trying to deal with this in regular mail. If that's the level of understanding it takes to write a CD we are all in trouble. Fortunately it's not, but you do make it clear this still isn't for Aunt Minnie.


From: Steve Setzer <setzer@backfence.net> Re; Java --impact of syntax

Jonathan Sturm's point is quite well taken, I think. There's an additional psychological element to Java--its syntax. Much of the syntax of Java is deliberately lifted from C++, despite the fact that Java's core design owes much to Objective-C and Eiffel, and very little to C++.

Many mediocre programmers are afraid of learning new syntaxes. So, Java succeeds among programmers where its ancestors failed because it "looks like" C++; unfortunately, that similarity leads the same d**ned lazy C/C++ programmers to apply their usual misbegotten programming techniques.

Visual Basic is an excellent toolkit built on a useful language. So are Delphi, MetaCard and WebObjects, albeit for wildly different purposes. Yet many programmers look down on all four, because "It doesn't look like C." Yeesh.

Steve Setzer

Thanks.

And from a reader who need help:

Dear Sir,

I enjoyed your recent article on www.byte.com about installing the Corel distribution of Linux. I was especially intrigued about your difficulties with printing over your network. I, too, am unable to print over my LAN and in spite of some expert help, am still unable to do so.

I have two computers, a server i386 running Mandrake 6.0 and a Sun SparcStation 20 running RedHat 6.0. I cannot print anything from SS20 to the i386, even though the i386 can print perfectly fine. Some of the same symptoms that you encountered I see- tcpdump shows the SS20 and the i386 talking to one another, but nothing prints and the print job winds up in its spool on the SS20. Also on the SS20, the status file says it is waiting for the queue to be enabled on the i386.

I tried setting the domain name on the i386 to "localdomain" as you described in your article. It still doesn't work. Previously, I have been messing around with the lpr commands. I've only had partial success-in that it seemed that shutting down the computers forced them to flush jobs and the test page printed out.

I realize you are not supervising a newsgroup, but RedHat support "doesn't support networking issues" and I am getting desperate. If you or anyone you know can help me, I would be very grateful. Thank you.

John Zbesko  johnzbesko@mediaone.net

 

 

TOP

 

This week:

Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday

TOP

Wednesday, December 22, 1999

Moore's Law for Digital media seems conservative. http://www.theregister.co.uk/991130-000011.html 

Is an article that describes a CD-sized disc capable of holding up to 140 GB of information.

That should hold a full digital projection movie.

Paul Wirtz

Fascinating. Thanks.


The language debates resume...

Jerry's input: I have always known that the real purpose of C is to employ C programmers; since the number of times any C code has been transferred successfully from one programming team to another is very nearly nil, if you do your project in C you will have a job for life or at least the life of the project.

Jack's Input: Well, let me give you my insight on this idea.

When I worked at Paradyne (www.paradyne.com) I was part of a database support group for the Operations / Management department.

The embedded applications department needed a database to manage the jobs going through their department, but rather than ask my department to do it, this group of C++ programmers decided to develop the database in (guess what?) C++...

Upshot was, in the time it too FOUR C++ programmmers to develop a database that _sort_ of worked, (atually it never did, well) I had developed three similar inventory / work management packages for 300 to 400 work stations on the shop floor in Clipper. (see http://www.ca.com/products/clipper.htm )

I work in Visual FoxPro, nowadays, which is a much faster language to develop in, and to run...<grin>

My final assessment is; C++ is great for SOME small jobs in which speed of execution is of the _essence_, like embedded apps, drivers, etc, but for REALLY big jobs... it's insane to use.

But then we have:

Hi Jerry

Regarding the comments made by Mr Jonathan Sturm:

> I have a friend who programs in Java. He's a Sun Solaris fan for > almost everything, but programs Java for his clients on an NT box > using Visual J++. I asked him why and his response was quite > intriguing. He says it takes at least 50 times longer to program > in Java, so he makes 50 times as much money. An application he > could knock up in a day or so using Visual Basic will take many > weeks in Java. He says he can get away with this because of > Sun's hype about Java and anti MS sentiment.

That sounds a little far-fetched to me. While it's quite likely that an experienced VB programmer can develop more quickly in VB than Java (particularly for small to medium sized projects), a factor or 50:1 seems quite unlikely.

Similarly, notwithstanding the Dilbertesque view of managers, most project managers could spot an price blowout of 5000%.

> I asked why use the MS product when he could be using something > from Sun.

Sun only provide a bare-bones development environment. Microsoft provide a rich graphical environment, allowing you to graphically design screens and drag-and-drop components onto the application. It is much quicker and easier than Sun's environment.

Many other vendors, including Borland, Sybase, Symantec and IBM provide similar graphical environments (IDEs), and all are very easy to use. The question of "which is the most productive" comes largely down to a matter of taste.

> Was he programming for Windows only and looking for the extra > speed from the Win API hooks? No, he said that Visual J++ was > a much more efficient environment to program in and that MS > making it harder to program "run anywhere" Java apps was total BS.

Its easy to say that, but its simply not true.

The main productivity advantage of the IDEs is the use of component palettes. By default, Microsoft's component palettes all contain Windows-specific components. While I'm sure it's possible to reconfigure the Microsoft IDE, it's not trivial (and the MS documentation doesn't help).

For more info on this you could check out my review for Australian Personal Computer magazine at http://www.apcmag.com/ (called "tool time").

If you don't believe me, try some other authors, for example PCWEEK: http://www5.zdnet.com/products/content/wins/0606/310599.html

> He said that VJ++ was so much better than other Java development > tools, that it is much quicker to develop in.

He is, of course, welcome to his opinion - however there it is far from unanimously accepted by developers.

Like all programming languages, VB and Java have their strengths and weaknesses.

VB is great for prototyping. It's fine for transaction-based applications - enter the data, save in a database, produce queries and reports. It's cumbersome for complex data manipulation or number crunching.

Java is good for network-based apps with good support for CORBA and other network protocols. Opinions are divided about it's ease of use, but it's still too slow for real-time apps.

My personal prejudice is for C++, but with a few riders. Its possible to write disciplined, maintainable C++, but it's much easier to write spagetti. I am considerably more productive in C++ than VB or Java, because I have spent many years learning how to program. It's not a language for somebody who wants to tinker. Either spend a long time becoming very good at C++, or stay well away from it.

C is very powerful for low-level (close to the machine) programming. It's particularly useful for small (embedded) computers, where memory and CPU are at a premium.

Fortran 90 is much maligned, but it has some very powerful features for number crunching, particularly for matrix manipulation.

Lisp, Smalltalk, Eiffel and many other languages are used happily in particular applications for which they are well suited. Not all tools are suitable for all jobs.

-Michael Smith emmenjay@zip.com.au Emmenjay Consulting Pty Ltd http://www.zip.com.au/~emmenjay/

Just when you've made it foolproof, along comes a better class of fool.


Greetings from St. Marys, Ontario, Canada

I've read you on/off/on for many, many moons -- first time writing.

In coming across your digital photos at http://jerrypournelle.com/chaospics.html#more1 , and in seeing that they were dark (typical for digital cameras, as I'm finding out), I thought fiddle around with their quality.

That was fine and dandy. But then I noticed that two or the images could possibly be stitched together.

Well, what do you know!? The result was better than expected! Instead of emailing you the image (plus going into more detail here), I invite you to visit http://stonetown.com/jp

And, of course, you're more than welcome to do whatever you please with this stuff; I can give back a little 'something'. :)) Many thanks for the countless articles you've put together and published -- yer one of my better reads. Best regards,

Alan

Alan Powell [aka webmason@stonetown.com] Discover St. Marys, Ontario, Canada ... 'THE STONETOWN' http://stonetown.com/stmarys

I must say that looks VERY nice. Thanks! (This came in weeks ago but got lost...)


 

Hi Jerry,

here's a feature I just discovered and found to be very useful for me: If you enable "show HTML tags" ( <Ctrl> + </> or menu 'View'->'Show tags') you can edit the tag properties by double-clicking on them.

Not sure if you could do this with Frontpage 98 but with FP2k I find this a major timesaver because in complex pages right-click is slower and sometimes ambiguous (does it refer to frameset, page, table or text inside that table, all objects being in the scope of the cursor at that moment?).

I went back to Frontpage after using GoLive 4 and Dreamweaver 3, BTW. Other products have shortcommings as well ...

Regards,

Moritz

That does in fact work well, and I had not known about it. There are many features in FP 2000 worth working with if only I knew about them...

 

 

 

TOP

 

 

 

This week:

Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday

read book now

TOP

Thursday, December 23, 1999

All the discussion about if digital projection or improvements to film handling can result in superior performance of a movie are missing the point.

It doesn't matter how nice the show is if you have to view it in crap.

In this neck of the woods, Melbourne, Florida, the problem isn't so much projection (although it always needs to be looked after), as it is the venue where you have to go view the movie.

Multiplex cinemas may get greedy and chain prints, but only after they stop cleaning the rest rooms.

Uncomfortable seating that's many times also dirty, rude patrons and employees, unswept floors sticky with spilled drinks, overpriced snacks, and more make me want to stay home or do something else.

I no longer go to the local theaters (one is less than a mile from here) - opting instead to drive to Vero Beach where there's a new 24-plex that hasn't decayed much yet. The same chain operates a set of 6-plexes on Merritt Island that's closer in distance, but takes longer to get to.

If I go to Vero, it's about $5 to get into the movie, plus about another $4-5 in gas (60 miles at 20 mpg). And there's still the overpriced snacks, rude patron and driving time issues. So I'm only going when I have nothing else going on or if it's a really big movie.

And I now have a DVD player. It cost me about the same money to go see "Iron Giant" in the theater as it did to buy it at Sam's Club last week ($16).

It's like the movie ("Other People's Money" - I think) where Danny DeVito had a speech about buggywhip manufacturers - the last movie projection system was fantastic, but no one wanted to see it.

Regards,

Bill Newkirk wnewkirk@iu.net

My wife and I still enjoy an evening out at the movies, and fortunately Hollywood still has many nice theaters, although the best place to go (other than El Capitan) is one of the three major theaters in beautiful downtown Burbank. Thanks to city manager Bud Ovram there is good parking, and the theaters are pleasant. What's more, although the theater plazas teem with teenage kids, they are POLITE, and even excuse themselves if they get in the way. But it certainly the case that DVD disks now cost not a lot more than an evening at the movies; I am sure I paid less for The Matrix than we spent to go see it.


I get requests similar to this at least monthly:

Jerry-

If you could hand someone a list of twenty five books that you felt a person must have read to be a well read and well rounded person in our society, what would they be? These can include books on science, history, religion, math, whatever _you_ think a person should read.

Thank you for taking the time to read this, and have a great holiday!

-Ryan Greene

The question is one of allocation of resources and credibility. I am not a professor of western culture, only an educated man. I do try to include works of broad interest in my "book of the month" at the end of each of my columns. 

I confess I have sometimes thought of making an "essential books" list, and now and again I even put some works I think worthy of attention up at this site (Lays of Ancient Rome, Hound of Heaven, come to mind), but I don't know whether it would be a good allocation of resources to try to compile an essential list which in any event would only be one more such list among many. Also, many of the works I think important just for getting started are long out of print... We'll see. Thanks for asking.

 

TOP

 

 

 

This week:

Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday

TOP

Friday,

 

 

 

TOP

 

 

This week:

Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday

TOP

Saturday,

 

 

 

TOP

 

 

This week:

Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday

read book now

TOP

Sunday, December 26, 1999

The future of the personal computer

Jerry

Let me say, that I am a computer user who prefers to select the software and hardware that meets my particular needs. I often get irritated when I am forced to use Microsoft applications that do not meet my requirements simply because an organization has standardized on these products. I would willingly concede MS 80% of the market knowing that their software meets the needs of this percentage of the population, even though other choices do equally well or better. It is, however, untrue that MS software meets 100% of user's needs, but it does make it difficult for alternatives to exist and remain viable.

I realize my comments will be construed as the rantings of a malcontent, but I would like to see computer columnists do a test that was focused on what knowledge users do to produce, revise, and polish a thesis or a legal document in Word, Word Perfect, and Word Pro. This kind of test would quickly settle the issue of which application is really more productive on a life cycle basis. It would be more relevant than testing the other applications against whatever features MS currently thinks are relevant.

I personally do not know how the current technical challenge between Windows and Linux will finish. It could easily be another OS2 scenario, or it could be a rout for MS. I think the outcome will hinge on unpredictable events. Thus, I find it somewhat amusing that MS enthusiasts look to the size of the company and the complexity of the product as a strength. MS is clearly in the sweet spot where IBM once existed, but look at what happened there.

The weakness of MS is the unnecessary complexity of its products. By attempting to make the average users task as easy as possible, they have made it extremely difficult to maintain an installation or use if efficiently in a life-cycle context. In my experience organizations that once had a few people looking after there IT now employ legions of MCSEs to keep the technology running. Yes, the technology does more now than it did then, but the question is whether or not a competitor can do as much or more with less. When you are talking large networks, it doesn't matter whether an OS is a little more difficult to build if it is easier to maintain and more reliable to operate. (Rule of thumb, front-end costs are less than 20% of lifecycle costs).

Have you ever wondered why the only segment of the market that has held on to Word Perfect is the legal profession. It is because WP allows you to produce complex documents that meet the exacting legal standards with far less effort than Word. I have used both Word and Word Perfect since they came on the market. While at some point it may have been far easier to create simple text with Word, the fact of the matter is that it can be a nightmare to edit a complex document created in Word. Word 97 table features were absolutely primitive compared to Word Perfect. I recently spent about three hours trying to stabilize section headers, section page numbering, paragraph numbering and the table of contents in a fifty page document I prepared with Word. I had to edit text to make the page-breaking more attractive because Word does not accommodate local changes like line spacing adjustments or font size or other parameters very easily. The client's requirements did not quite match any of the templates built into Word. Thus, for example, because non-standard styles had to be accommodated in the table of contents, each time I re-generated the table, I found that Word had reverted to its defaults. Just try and do something that the program does not support without writing a new template to do it.

Ian McCreath iandmccreath@home.com

In fact, most law offices now use Word, and even church bulletins that used to be done in Word Perfect are being done with Office which includes Publisher. At the moment the momentum is on the side of Office, not Word Perfect. Corel Office for Linux may slow that trend. I don't know anyone who thinks it will seriously reverse, at least for a while. Perhaps so.

 

Dear Dr. Pournelle:

I hope that you and yours have had a Merry Christmas. In looking over the news today, I read about the repair of the Hubble Space Telescope. What caught my eye was the statement that the 386 onboard computer was replaced by a 486. At the time the Hubble was launched, I could see a 386 class processor. But now? Of course, my confusion may be due to ignorance on my part and there are excellent reasons to make such a choice. Assuming however that this is an example of the "better, cheaper, faster" mantra of NASA, I have to question just what grade of spacecraft and electronics we are sending into space these days. This is especially true when the latest Mars flop is taken into account.

When I was a boy watching the space race take shape, I was proud of the fact that our launches went well (at least after the fire at the Cape) and the US did what we set out to do. Later in life I was amazed at the differences between the Russian and American spacecraft in terms of fit and finish. I learned that Russian engineers used "off the shelf" components whereas American philosophy was to design from the ground up. From what I am reading, it looks like NASA is moving to an "off the shelf" mode. What concerns me is that NASA appears to be shopping off the clearance shelf from Fry's.

This would not be so bad, except that at the same time one reads that NASA is doing its best to kill free enterprise space capability at the same time.

Am I reading too much into recent news? Or is this a problem?

Richard Cartwright rcartwr@attglobal.net

Rad hardening is tougher than it looks, and the 486 is probably more than good enough; I know that when I introduced the 386 to the Lowell Observatory equipment it was plenty powerful for what we wanted to do then...

Hi Mr. Pournelle.

I just read your piece about Linux in Byte Online. It was good read as usual.

I was just wondering: have you tried Sun's StarOffice for Linux? It can read and write MS-Office 97 files and has all the apps you can possibly want from an office suite. (www.sun.com/staroffice). Plus it's free (for now, at least).

I am currently starting a process to just leave the Windows word for good (specially now that a lot of games are being released for Linux too).

Thanks.

Wanderley

I have mentioned Star Office, which has a number of decent features. And yes, there are games being ported to Linux. I wonder if anyone has done Master of Orion (MOO) for Linux? It's still one of the best around...

 

 

  TOP

 

 

birdline.gif (1428 bytes)