Dr. Jerry Pournelle

Email Me

Why not subscribe now?

Chaos Manor Subscribe Now

Useful Link(s)...


Hosting by

Powered by Apache

Computing At Chaos Manor:
April 24, 2007

The User's Column, April 2007
Column 321, part 4
Jerry Pournelle jerryp@jerrypournelle.com
Copyright 2007 Jerry E. Pournelle, Ph.D.

Continued from last week...

I have some advice: if you are planning on getting a new computer with Vista, put off your decision for a while. Unless your need is urgent, make do with what you have. You'll be happier.

I know. In past columns I have said don't upgrade to Vista from an older machine, but I didn't caution you not to buy a new all-up system with Vista. Now I am saying don't buy a Vista system at all; at least not just yet.

What's Wrong with Vista?

It's not so much what's wrong with Vista, as what's not right: that is, while parts of Vista are fun (if you have a sufficiently powerful machine), and the search function is not bad, there's really no reason why you must have Vista. In theory Vista security is better than XP security, but in practice the real difference is User Account Control.

Now if what you do with computers is run a small set of programs, and you don't do much Internet surfing, and you don't install new programs, and you don't often open new programs, then you may be able to operate with User Access Control turned on without going mad. I say may be able: I have two systems with Vista now, and my record for running Vista with UAC turned on is about seven hours. I didn't manage a full day with UAC before I turned it off with extreme prejudice and many curses. Your mileage may vary, but I'll bet it won't vary by much.

Another part of Vista Security has to do with internal networking. In my case Vista is so secure it won't connect to some of the other machines in my local net. It can see them, but when I try to connect to them I am told I am not allowed to do that. Note that the system I cannot connect to with Vista is part of my network, and all my XP machines can connect to it. Moreover, Roxanne, the Vista system, can connect to most of my other XP systems, and both my Vista systems can see and connect to each other. When it comes to Wendy, an Intel Pentium 4 that was formerly one of my main machines, the two Vista machines see her just fine: but attempts to connect fail with an unfixable error. I am told to see my network administrator.

Incidentally, Wendy can connect to the two Vista machines. I can even map Vista machine drives to a drive letter in Wendy and use xcopy to transfer files. I just can't go the other way.

Now I make no doubt that I will find a way to get my network set up properly. I am certain it will involve turning off some security features, possibly some in Wendy as well as in the Vista machines. Given that I have to turn off most of Vista's security features in order to get any work done, it's hard to say that Vista is more secure than XP.

Search under XP

As to the improved search features in Vista, I don't know how to use them properly, because I don't find Vista search any better than the Microsoft Desktop Search add-on program for XP. When you install the Desktop Search XP will have two search and find methods: the old, slow, unindexed, clunky search you find in the START menu, and the very rapid indexed Search Desktop you will find down in the tool bar at the bottom of the screen. That latter search feature is nearly instantaneous, so it would be very hard for Vista to do it better. It's also a lot more convenient: the Vista search function is harder to open, at least for me.

Peter Glaskowsky tells me that all I have to do is hit the "Windows" key and start typing. OK, I tried that: this column has a review of Max Boot's latest book, and that was written a week ago and is present in a file called "2007 April Winding.doc". Over on my XP machine, the instant I type "Boot" in the desktop search window I am given a list of folders that includes the one I have in mind. Here on the Vista system I tapped the Windows key, type Boot, and get an irrelevant folder. Just telling Vista to look in all the folders on the C: Drive takes a lot longer, and the search takes forever. Hitting the Windows Key and typing "Boot's" gets the instant result that "No items match your search." Over on the XP machine I am instantly shown the files that contain that word.

Another example: if you use Microsoft Word, try using Vista search to find "custom.dic", the default dictionary that words get added to. Now go to an XP machine with Microsoft Desktop Search and use the old-fashioned "Find" using the search companion. You'll learn a lot about finding hidden files and folders – neither search system indexes those by default – but you'll also learn something about the speed of each search system. (And yes, I know, you can use Word itself to find the dictionaries, but that's not the point.) Now try finding all the .pst files on the two systems. You will be enlightened.

I kept fooling around, and I found that by default the Vista indexing system doesn't include the primary drive, or very much else outside "MY DOCUMENTS". Since I don't keep much in "MY DOCUMENTS" (as dumb a concept as I can imagine) it wasn't finding things for me. I have corrected that, and it certainly works much better now. I'll report on it next month. Meanwhile I can only say that Vista's search indexing defaults make no sense whatever; if you're using Vista go find them and change them. Once you do that, things improve enormously. Pity it took so long for me to learn that.

Even so, I very much miss the Microsoft Desktop Search Bar. I wish I had it on the Vista machine. I am slowly learning what I have been doing wrong in Vista, but my point is, why must I be told how to use something that should be obvious?

If you are using XP and don't have the Vista-like search capability installed, you can find out how to do that at this link from TechRepublic.

Also, the Microsoft Desktop Search is excellent. The indexing can take a while, but once it's done, the search speed is wonderful, and it's very intuitive. Install the Desktop Search or the Vista search capability on your XP system and you will have one less reason to get a Vista system.

Windows Vista Secrets

If you're still inclined to get Vista, do yourself a favor: go buy Windows Vista Secrets (Wiley) by Brian Livingston and Paul Thurrot (reviewed later in this column) and at least thumb through it. Brian is an enthusiast, and he may convert you to wanting Vista; that's a chance you have to take. The book will also alert you to just how complicated a decision you are about to make. And if you do install Vista, I guarantee you will be glad you have this book.

Why Wait?

Assuming you're staying with Windows – that's a subject we can get into next week – you're going to learn Vista some day, so why wait? Why not go buy a powerful Vista system and get started?

The main reason is that it may not work. For reasons I don't understand, peripheral manufacturers have had over a year to develop drivers for Vista, but most haven't done it. Walter Mossberg's experience is typical: he reports in his Wall Street Journal column "the only built-in Vista printer driver I can find for my printer doesn't allow the two-sided printing I can do with Windows XP and Apple Macintosh computers."

Uncle Walt's experience is typical. Many of my readers report that they can't find decent drivers for much of their hardware. Those will be coming Real Soon Now, but they aren't here yet. Uncle Walt's advice is "if you're not desperate, you might wait another six months or so for the software and hardware to catch up – and for Microsoft to do some bug fixes."

Skating the Fine Line

Microsoft didn't always have an effective monopoly on PC Operating Systems. I won't repeat the history of Microsoft, Digital Research, IBM, and early PC DOS programs. The key competition was between Microsoft Windows and IBM's OS/2. Microsoft won that in the market place, with the active assistance of IBM. I've told the story before: at one Spring COMDEX, IBM breathlessly announced that you could buy their Software Developers Kit – a CD that enabled you to write drivers for OS/2 – at the greatly discounted price of $600. Microsoft, at that same COMDEX, was dropping CD's of their SDK from airplanes. It wasn't possible to walk within fifty feet of the Microsoft booth without having an SDK thrust into your hands.

Microsoft also shipped their products fast. If they worked at all with existing hardware, they'd work a lot better with the hardware people were buying. Microsoft understood Moore's Law. Ship it, the hardware will bail you out.

IBM came from a more leisured tradition, in which hardware upgrades were rare, and what was important was to get things right. Don't ship it until you're sure it works. By the time the IBM Executives figured out that this wasn't a winning approach, it was too late.

Microsoft won that competition by shipping early and sending out the fixes. Ever since, pundits have been saying "Darned right Microsoft has a quality control department. It's called their early customers." The standard advice most of us have given for years has been, "don't buy the first release of a Microsoft product. Wait for Service Pack 1 before you buy."

So many people took that advice that Microsoft expanded beta testing to include a great many people, and made bug reporting a lot easier. That, they thought, ought to do it: when the product goes gold, enough people will have been using it for long enough that we will have the bugs fixed – and the Internet makes it a lot easier to send out those bug fixes.

Many people, including me, thought that the extensive beta tests would be enough. Surely, with all those users pounding on Vista, it would be in pretty good shape before it went gold.

We were wrong. Yes, Vista works pretty well. It seldom crashes, and when it's working it works well; but there aren't enough drivers, a lot of legacy hardware won't work properly, and, when you come right down to it, one reason Vista is as stable as it is, is because Microsoft stripped so many advanced features out of the shipping product that what's left is more cosmetic than exciting new capabilities. And for heaven's sake, if you just have to have a new machine with Vista on it, don't get one with Vista Home Basic. That one won't even have the cosmetic features.

The bottom line is, Vista will be the best Windows yet, but if you don't desperately need a new Windows system, hold off a while. I use Microsoft applications every day, and I find my XP systems more than good enough on all my equipment including my (old but I still think of it as the "new" laptop) T42p ThinkPad, and even Lisabetta, my almost ancient HP TabletPC. Now true enough: I have added extra memory and larger disk drives to both those machines, and without those I'd be a lot less happy.

Of course you may have noticed that Dell, among other computer vendors, is offering machines with XP again. This is another alternative. Get a good Intel Core 2 Duo system or the AMD equivalent with two gigabytes of memory and XP. You'll love it, and you can upgrade that to Vista when SP-1 comes out.

A final alternative for those who need new Windows capable hardware right now is to look at the new Apple systems with Parallels, Boot Camp, or VirtualPC. If you go that route, there are many options, depending on what you intend to do both now and in future. I'll discuss Apple system options another time. Meanwhile, if your present XP system is good enough, hang on for a while. Contemplate adding memory.

Winding Down

The first book of the month is The Crazyladies of Pearl Street by Trevanian. It is his last work. Trevanian died last year.

I was unaware of this book, which is strange. Trevanian was the pen name of Professor Rod Whittaker, an old friend of mine from Seattle days. I was the company manager of his Master's Thesis play at the University Playhouse, and we were fairly close friends when he was in graduate school and I was an engineer at Boeing. I left Seattle for California in 1964, and lost track of Rod and Diane; I never saw either of them again.

Rod became a Professor and Department Head at the University of Texas at Austin during the time when I was on the Board of Visitors of the Department of Mathematics at that institution; but oddly enough I never knew that, and there was no reason for him to know of my association with the UT math department. I wrote him several letters in care of the last address I had for him, but none ever got to him. I certainly did not know he had become a famous author.

Rod wrote as "Trevanian", and the identity behind the pen name was a fairly carefully kept secret for many years. Eventually I learned that identity and was in correspondence with Rod and Diane, but it was toward the end of his life. Diane and I still exchange rare emails.

The Crazyladies of Pearl Street is not quite an autobiography. I never knew many of the details of Rod Whittaker's upbringing, but I know a few of them, enough to know that the boy in this novel is not quite him, although they share many experiences. Think of Jean-Luc (the boy in the novel) as a younger Trevanian, not the younger Professor Rod Whittaker, and you'll be close to correct. For all that it is real enough. This is neither novel nor autobiography: it is a remarkable work that embeds a young boy in his times, as remembered by the mature writer that boy became. As you would expect, the writing is precise and exact. Rod was a born storyteller.

For those interested in America from 1920 to 1940, there are two major works by Frederick Lewis Allen: Only Yesterday, and Since Yesterday. The first is out of print but is available on line. It covers the period until 1930. The second, Since Yesterday, covers the 1930's, and is available from Amazon. These books cover the history of the time, and give a certain amount of understanding; but to really understand those times of the Great Depression, you need to live them. A few books accomplish that. One is Fred Pohl's The Way the Future Was. Fred lived in New York and saw the big events.

Rod Whitaker was in Albany, and saw the world of the Depression and World War II from the streets through a boy's eyes. Trevanian's Crazyladies makes you live those times. It's very readable. It's also a work of genius.

The second book of the month is Max Boot's worthwhile but flawed War Made New. This is a combination history of technological war and an attempt to present a theory of high tech war. The history is well done in places, and oddly careless in others. For example, on page 30 Boot says of Queen Elizabeth I, "Papist power was a direct threat to Elizabeth herself, because Catholics did not regard her as England's rightful ruler and were constantly plotting, sometimes with the assistance of Spain, to depose her in favor of her Catholic sister, Mary." Of course this is nonsense: Elizabeth's older sister Mary, known to history as "Bloody Mary," had been married to Philip of Spain as well as Queen Regnant of England; it was only on Mary's death that Elizabeth became Queen. One presumes that Boot confuses Mary Tudor with Elizabeth's cousin Mary Queen of Scots, but aside from some plots, all discovered by Walsingham well before they were of much threat to the Queen, Mary Stuart was never a real threat. The only real threat was the Spanish Army, which, could it have crossed the twenty miles of the Channel, would have subdued England with no more trouble than William of Normandy had in 1066.

Boot's history of the Spanish Armada is conventional, and completely leaves out the Sea Beggars who actually prevented the Spanish Army in Flanders from crossing the Channel while the British Fleet was engaged with the Armada. For that story one should turn to Fletcher Pratt's Battles that Changed History, which is a much better (and far shorter) work, and indeed one of those books that everyone interested in Western civilization should read. Much of Boot's history is like that: it tells little that's not better told elsewhere, and doesn't add much to his case.

Boot's strength is in his later chapters, and one wonders why he left the early material about the Armada and 1588 in the book at all. It adds little. Fortunately, his later chapters on military technology in the modern world are considerably better and worth your attention. Military technology is indeed decisive, as Possony and I showed in our Strategy of Technology. No existing organized military force can stand up to the United States Army in battle; and it would be exceedingly difficult to impossible to create one able to do so. Boot understands that kind of war very well, and these sections of his book are important. Much of Rumsfeld's work in building a new model army was well done, and implemented important changes in the military bureaucracy.

What Boot does not do (and Rumsfeld never learned) is address the problem of how a modern combat army can fight insurgents. It is an old maxim of military science that good soldiers do not make good butchers. A combat army is generally not much use as an army of occupation; a fact known since Roman times if not before, and that the Israelis learn every week.

A democratic army of occupation is an even worse contradiction in terms. Occupation and reconstruction worked in Germany and Japan for several fairly complex reasons. Those factors do not apply now, but Boot is apparently not aware of this. The stark fact, though, is that America is governed by consent of the governed. Training an army to rule without the consent of the governed is usually not a very good thing for a democracy to do.

The Roman solution to the problem was to recruit stationary constabulary forces for police and occupation duty. These were often originally supplied by client states, but over time most of the constabulary was recruited locally from the occupied peoples. They stayed where they were recruited, and seldom moved far even to support actions in the next province. Meanwhile, the heavy infantry Legions stood ready to suppress invasions and to make certain that the auxiliaries did not attempt to become rulers. Legions might propose new Emperors, but the local auxiliaries never did. Given that Boot and the neoconservatives advocate US long term military commitments in the Middle East, it's surprising that not much of this is discussed.

Alas, few books do examine these questions. Building a constabulary occupation force is an entirely different proposition from building a decisive military force whose main mission is to break things and kill people. I can recommend Boot's book as one of several works to be read in understanding modern military questions. It should not be the only such book one reads.

The computer book of the month is by Brian Livingston and Paul Thurrott: Windows Vista Secrets, Wiley. We mentioned the manuscript version of this book before. The final product is better than the early review copies. If you are serious about using Vista, you will need this book; it's both handbook and instruction manual.

Incidentally, Livingston and former Windows Magazine (and prior to that, Byte Magazine) Editor Fred Langa have joined forces. After Fred left McGraw Hill's BYTE for CMP's Windows Mag, we considered him our greatest rival, and when CMP bought BYTE from McGraw-Hill we thought the two periodicals would merge with Fred as EIC of both. Instead, CMP shut both magazines down. Inexplicably they closed BYTE so fast that there was no staff left to return review equipment, and the same people had to be hired at consultant rates to finish the orderly closure. Then, later, BYTE was revived on-line, where it remains.

The second computer book of the month is Ellie Quigley's PHP and MySQL by Example, Prentice Hall . One of the cover blurbs proclaims it "The perfect introduction to PHP 5 and MySQL," but I wouldn't call it that; it's hardly introductory level. However, if you have any familiarity with PHP and MySQL, this book will take you a long way. It comes with a CDROM of all the examples. These cover almost everything you might want to do. If you work with these programs, you will need this book. That, of course, is the view of someone who doesn't do much programming now; the opinion comes from experience. When I did write programs, I found that examples were the best teaching tools I could find. If I had an example of program language that did more or less what I wanted to do, it was a lot easier to modify that than to start from scratch. If your programming skills are so good that you don't need examples, you can safely ignore this review.

The game of the month was Microsoft's Freelancer, but I can't say I spent much time playing it. It was fun, and if you have good eye-hand coordination you may be able to live through the missions in the story line. I suspect I could train myself to do that – it's not really harder than Privateer was – but I have been so busy working on fiction that I have had little time for games. Freelancer may not still be in print, but it's available, and is the closest thing to the old Wing Commander Privateer I have yet found. Alas, while I had no trouble installing Freelancer on XP machines including my T42p ThinkPad, I have been unable to get it working on my Vista games machine. In theory it ought to play under the XP emulation option, but in practice I can't get it to install.

I also had some fun playing about with Second Life. If you haven't looked into the Linden Labs world Second Life, you should. It doesn't cost anything unless you want it to, and there's just a lot going on in there. There's an extensive active space advocacy community among many other interest groups.

The movie of the month is The Last Mimzy. It's not as good as it could be, but it's worth seeing. The plot is loosely based on a short story entitled "Mimzy Were the Borogroves" by Lewis Padgett – the pen name of Henry Kuttner and his wife C. L. Moore. I never knew Kuttner but I was privileged to know Catherine, so I was bound to see this movie in any event.

It's very likely you'll also like Blades of Glory, the latest Will Ferrell farce. Roberta took me to that one evening after I had spent the day working on fiction, and it was a great relief from the Inferno I'd spent the day in.