User: suitti

Re:What to use that for

On: Introduction to 64-bit Computing and x86-64

I bet it rips mp3s like a motherscratcher.

Funny, I've been doing alot of mp3 stuff on my Pentium II of late. It encodes mp3s at about double real time, with notlame under Linux. That is, I can fire off a conversion, and immediately fire off xmms to play the file as it is converted. xmms will never catch up, because the conversion happens faster than real time. Now, a PII/350 is not a modern machine, and nearly everything on the market blows it's doors off. However, it isn't slow enough for me to be able to justify replacing.

Perhaps if I ran Windoze, where my machine would be largely useless until the program finishes, I'd change my mind.

On the other hand, I have an app that I wrote, which spends half it's time computing a has function. I use 64 bit integers (long longs in gcc) to generate better values than one gets with 32 bit integer arithmetic. On an Alpha (64 bit arch) this hash function screams. Currently, the hash function is used as an index into a bit array. I currently use a 512 MB bit array. I'd like to bump this up. Now $800 will buy me 8 GB RAM. But my x86 can't use it (mine is limited to 1 GB).

I could probably get a used Alpha, Sparc, etc., and run this. But I'd rather have compatibility with the x86 arch, if only to get access to the cheaper commodity RAM.

For my app, larger address space and larger physical RAM capabilities are more important than additional speed. This is for an app that runs about a half hour per run, but is often run continously for about a month at a time.

Numbers

On: A College Without Microsoft?

Macs come with Outlook and IE these days. They may have MS Office on them. These may be replacable with Word Perfect or OpenOffice and Netscape. This would only be required for new Macs - as existing software could be grandfathered. Macs can last a LONG time (mine is now 16 years old, and I still use it). Still, at least some of them should attempt migrations. I'd attempt doing it for all of them. The Outlook family members are virus/worm supporting platforms.

Existing PC's wouldn't have to be converted right away. Still, you'd want to migrate to Linux to get experience with the alternatives. If you have tons of Word documents, you should start exporting them to RTF, and experimenting with using them on new software right away. You DO run into issues.

Let's say that you use the $800K/year for new hardware, and hardware is turned over every 3 years. That gives you $2.4 million for your total hardware budget. For 7000 students, that's $342 per student. That's a pretty cheap system for each student. If you go with 4 years for hardware turnover (average) you get a much easier $457 per student. For that, you can get useable Linux systems from Walmart.

You might want to allocate some money to a few higher end systems. These would be put together from parts - which would save some money, get you exactly what you want, and still avoid the MS tax.

One outcome of this experiment would be a working cross platform word processing file format. That, IMO, would be worth the effort.

The price point might make you rethink individual systems. If they all have a network card, they don't all have to have CD drives or CD burners. These can be shared. A few might have scanners. They could all be used to back up each other's data. The network is the machine. This idea may make it hard to purchase Walmart systems, though.

Re:Professional sound? What about desktop sound?

On: Linux Audio Developers Conference

What is the problem? My Pentium II/350 with xmms can play mp3 using about 0.3% (1/300th) of my CPU. No skips, even when the load average is 10 (because I'm using 10 simultaneous instances of notlame to convert CDs to mp3). Even my old 386/33 could play sound (until it died - may it RIP). It ran Slackware. I wasn't doing mp3 back then, though.

My sound card is an ISA bus Sound Blaster 64. It seems to provide good sound.

Others have talked about how their IDE disk drives didn't do DMA, or how upgrading their kernel improved things. I'm using a stock Mandrake 8.5 beta install kernel. I customise it by adding modules. I used to recompile, but this time around, there was no compelling reason to do it.

Perhaps there's something else about your system that could be improved.

Re:Finally

On: Linux Audio Developers Conference

I have a Pentium II/350 with an ISA Sound Blaster card, Matrox AGP accelerated graphics card, 650 MB RAM, IDE disk, SCSI CDRW. It currently has Mandrake 8.5 beta installed. I'm using the install kernel.

Mandrake found my sound card and auto configured it. I run GomeCD or XMMS for sound. I just recently got an MP3 player for my car. I'm currently reading my CDs, converting them to MP3, and cutting disks for my car.

So, I pop a CD into the drive. Run cdparanoia to generate .wav files on disk. A script runs notlame to convert these to MP3. Since my CD drive can read CDs faster than my CPU can convert them, I background the notlame script runs, and I may have several running at the same time. I don't bother to nice them. While all this is going on, I use xmms (which was in the Mandrake distrib) and play directories of sounds. I might be looking for all-instrumental tracks for inclusion on a disk, so I copy those tracks to some destination. xmms seems to consume about .3% (1/300th) of the CPU. It never skips, even when the load average is ten, even when I'm copiing megabytes of data from one disk to another. And when I'm burning a CD, I don't even bother to stop running SETI@Home.

My point here is that if your Linux sound system skips, it isn't the CPU speed. My old 386/33 running Linux could play sound files without skipping (until it died).

In the mean time, my friends with Windoze running on GHz+ P4's have to stop everything when cutting CDs because buffer underruns mess up everything.

PC hardware is complicated. For every component, there are order of magnitude price ranges. Driver issues are often unknown until it's too late. Some of these are more important than others. Many systems are misconfigured. This may not be a Windoz/Linux issue. It may be more of a PC thing.

For example, I bought a sound card for $12 (5 years ago). It claimed that it could do CD quality sound. Maybe they meant 16 bit per channel 44 KHz (more or less) stereo. But the sound quality was terrible. Who knows, maybe the A/D was bad. Maybe it wasn't very compatible with the driver I used. Whatever. I replaced it.

I've never seen a Mac that does sound poorly. Maybe having a single hardware vendor isn't such a bad thing. The sound (22 KHz, 8 bit stereo) on my 16 year old Mac II is still great. How's the sound on your 286?

You might check out DeMudi, which is a Linux distrib devoted to sound.

Static is the way to go

On: Microsoft to End DLL Confusion

The physician takes an oath "to do no harm". I'm considering putting together a Linux distribution with everything compiled statically.

First, what are the advantages of DLLs?

• Less Memory Footprint
• Less Disk Footprint
• Global Security Fixes
• Use of third party binaries
• Plug ins

Less Memory Footprint

In Unix, when you have two instances of an application running, say, vi, the executable code between the two is automatically shared. The shared library gains you nothing. To gain memory footprint, you need to use the same shared library from two applications at the same time. For example, libc might be used by vi and cc.

However, if you compile statically, you bind in only the routines that are needed. For shared libraries, you need to have all routines available, since you don't know which of them are used. Now, your virtual memory system may notice that a shared libary page isn't used, and page it out. Yes, this requires additional run time execution time. The upshot is that you save memory only when you have enough different programs use the same shared library to overcome the overhead. I claim that this happens with libc, libX11, and not a whole lot else.

Less Disk Footprint

If you have 50 programs that use the same shared library, you can save some disk space becasue that libary code does not need to be duplicated that many times. However, shared libraries need to have the symbol information requried to perform the dynamic binding. The savings isn't that much.

In the old days, when an entire Unix distribution fit on a 150 MB tape, the libc shared library savings amounted to about 30%. You could get more reduction in size by using compression.

In fact, programs could be compressed on disk. The loader could decompress the image as it reads it into RAM. For slow disks, this may be faster than loading the uncompressed version into RAM. The down side here is that you then may not be able to use the original file on disk for virtual memory paging.

In any case, it's getting hard to get a disk drive under 20 GB. 30% overhead reduction for the most common shared library doesn't amount to much.

Global Security Fixes

So, your libzlib.so.5 has a bug. You whip up a quick fix, create a new libzlib.so.5, and drop it into your system. You've just fixed all of your libzlib dependent programs, right? In fact, you fixed programs you didn't even know used libzlib. You may also have broken programs that you didn't know use libzlib. And, short of testing every program on your system, you don't know. The more complex the patch, the more likely you are to have broken many things.

Quick. What is a utility which will tell you all the shared libraries that an application uses?

Use of third party binaries

Third party binaries can just as easily be distributed in source form or in a library that is statically bindable. Static binding is preferable, since you are unlikely to use a large fraction of a kitchen sink shared library - where the authors have no idea how the programmers will use it. Source is preferable, since the documentation rarely specifies enough semantic detail to allow proper use.

Plug ins

OK. Your application is Apache, and you want some real flexibility. If Apache is compiled so that modules can be loaded at run time, then the administrator can add the new module and turn on it's use in the configuration. This doesn't save any RAM or disk, but it may allow the admin to change a line of config, restart the web server, and start using some new feature.

For Apache, the admin can also recompile with the new module compiled statically. I've done it both ways. My measurements show a small run time advantage to static compilation.

Granted, if you can't recompile IIS, then DLLs will give you the same flexibility in exchange for a small performance penalty.

The Dark Side of Shared Libraries

If you compile your application statically, then upgrade your OS, you can copy the old application to the new OS, and it just runs.

If your app has shared libraries, you have to track them down on your old OS, and copy them to your new OS. If you make a mistake, and copy your old libc.so over your new one, you run the risk of trashing every program on your new system. Brilliant.

Take netscape as an example. It comes installed in it's own /usr/local/lib directory tree. In /usr/local/bin, netscape is a script which sets up the shared library search path to include the libraries that netscape needs, then runs the binary. This introduces script overhead and shell dependencies on a complicated package. And, when you upgrade your OS, you still need to find the old libc.so and copy it forward.

RPMs

Many seem to think that RPMs solve all these problems. However, many packages have bugs in their dependencies, etc. Many RPMs use different versions of the same shared libraries. I find that I have to override the dependencies to get stuff to install. Often, the requried package IS installed. Not just once in awhile. Much of the time. The difference between theory and practice is that, in theory, they are the same.

Conclusions

Shared libaries seldom save RAM or disk space. The problem with using them to fix bugs globally is that you don't know what you fixed, or even if you broke some things. Third party binaries should invariably be statically linked. In an open source environment, plug ins are not strictly needed. Shared libaries make OS upgrades more painful.

So, what I'd like is a Linux distribution with no shared libaries. The compiler, gcc, would be configured to compile statically by default. Then, after some years of running the system in production, and after adding hundreds of applications to it, I'd be able to upgrade to a new distribution without having to recompile or do the shared library search.

Re:DLL vs static libs

On: Microsoft to End DLL Confusion

If you lie to the system, and say, "This new libzlib.so is libzlib.so.5", you still run the risk of breaking all the apps. The only way to know that they all still work is to test them all.

So, you change the shared library. What programs did you affect?

Re:Well of course

On: What High End Unix Features are Missing from Linux?

'awk' as released by AT&T was carried forth, for many generations of UNIX versions, and every vendor of every variant invariably failed to fix any of the common bugs.

I worked for Interactive Systems. We had our own SVr3 for x86, and did a number of other ports. We offered AT&T our bug fixes. They didn't want them.

There was no CVS then. If you wanted to fix the bugs, you had to apply the patches by hand. We had over a thousand patches that I was aware of. If AT&T incorporated our patches, it would cut down our porting time (and everyone else's) and increase the quality of Unix everywhere.

There was some bug in 'vi' having to do with slow serial lines and keyboard function keys that I fixed at least twice... This bug showed up in several other ports.

By contrast, if I found a bug in the PDP-11 C compiler and built a patch, Keith would get my patch into the next BSD release, which usually would come out in a couple weeks.

Proprietary systems make progress nearly impossible.

Bill Gates has basically said, it's not a bug if unless it causes customers not to buy it. I'm hoping that the Open Source model will show Bill just how short sighted this policy is.

Competition

On: China Wants To Establish Moon Mining

'Keeping technological lead' worked for NASA before. In fact, if NASA could donate 10% of it's budget to China, it could probably get Congress to double it's budget!

Sample Return

On: China Wants To Establish Moon Mining

I hope that china does a robotic sample return mission before sending people there. It still bothers me that NASA sent people there first.

Re:Not feasible

On: China Wants To Establish Moon Mining

It's worse than that. If there were rocks of pure platinum on the moon, it would still cost more to get it than it'd be worth on Earth.

Asteroids may be different. You don't have to get the stuff up out of the Moon's gravity well. You might be able to cheaply divert an asteroid to low Earth orbit (using solar sails or something), and even deorbit it for a crash landing in a big desert. For a big rock, you might have enough stuff to make it worthwhile.

You might mine an asteroid to build space craft in space. At this point, you'd be reducing the cost of getting a space craft built and in space.

We landed a probe on Eros. Were it designed to scrape He3 from the surface, and if we were ready to use it, it might be worthwhile.

Re:The future? Just like the past should be...

On: More on Columbia

Now that carbon nanotubes have been produced, and have shown to have a suffient strength to weight ratio, and are being produced in macroscopic lengths, the base technology for a space elevator appear within sight. It's strange that no contest was required. By contrast, enormous money has been dumped into hypersonic scram jet engines, and we don't have a working engine yet. Prize or no prize, without a base of technology, it will remain a dream.

Still, offering prizes might work, but the goals need to be revised. At a failure per 50, going to space with humans should absolutely require humans in space. (Yes, don't repair the HST - send up another one). Robots can do anything humans can do now in space cheaper. From a safety stand point, note that Galileo will be flown to destruction at Jupiter as the natural end of the program. Got a one-way mission? Send a robot!

The only goal I can think of worth risking humans is building self sustaining colonies in space (L5 Society), on the Moon, or Mars. To achieve these goals, you need to radically cut the cost going to orbit (space elevator) or radically cut the mass you need to put into orbit. You need artificial gravity. You need a self sustaining biosphere. For Mars, you need serious radiation protection for the trip. (To go to the moon, you need to be lucky - which we were for Apollo). (Once there, you can go underground for radiation protection on either the Moon or Mars.)

ISS and the shuttle are not getting us to these goals. There is no progress, so these programs should be axed or radically revised.

A proposal for ISS was to use inflatable structures. This would radically cut the weight (and cost) of the station. We went with high cost without progress.

Artificial gravity can be achieved by spinning. Perhaps some booster upper stage could be tethered to a station, and the whole thing spun up. (You also need to allow docking to other space craft, and deal with emergency exit - all perhaps at a slow moving center.)

Biosphere 2 (on the ground) showed that building a working biosphere is difficult. We need to keep at it.

We have now soft landed a probe on a minor planet (Eros). One way to reduce mass required in orbit is to make use of material that's already there. There's lots of material to use in Earth crossing orbits. But we haven't studied these bodies enough to know how to use it.

$5 billion for 3 years of a space station would make ISS look extremely expensive. We're at, what, $95 billion for ISS? I remember when $25 billion was the cap.

The carrot has already been dangled for EELV. We now have successful launches of multiple boosters. Oddly, it has contributed to a world wide booster market glut. This may be a good thing. But EELV is, by definition, not a radical change in booster design. SSTO (single stage to orbit), etc., are radical.

The Mars Pathfinder program showed that a cheap robot sending pictures home can be a good public relations tool, while achieving real science.

Half of the probes to Mars have failed so far. Reliability needs to be demonstrated before sending humans there.

When do you stop using a successful resusable launcher? When it fails? Will the shuttle program end when we run out of shuttles? The shuttle's life has already been extended...

spin

On: Intel: No Rush to 64-bit Desktop

Intel says they're in no hurry, but they've been working on 64 bit processors for awhile. The Itanium sounds like it ought to be a performer, but when they produce silicon, the benchmarks haven't shown it. Sounds like spin to me.

I'd like to see one of two systems. Either provide backward compatibility - like AMD with it's 64 bit extensions, or start with a clean slate and produce a performer - like Digital's Alpha.

The advantage of a 64 bit AMD is that the most used architecture can migrate without dropping everything. My PII can still run DOS binaries that ran on my 8088. This is a GOOD thing. Even running Linux, I don't want to recompile all my apps, if I don't have to. If this were the case, I might have gotten a Power PC already.

The advantage that the Alpha has is speed, and there is only one kernel systems calls interface - 64 bits. For example, there's no lseek() and lseek64() on the Alpha. (For the history buf, first there was seek() for 16 bits, then lseek() for 32 bits. We've been here before. Now we have the off_t typedef, so it should be easier to simply change it to be 64 bits... Yet some have added off64_t, in the name of backwards compatibility.)

Itanium may have the clean break (or it may not), but where's the speed? I'm not switching without something.

Digital's Alpha is at least the third attempt that Digital made before getting a RISC system to perform. The Power architecture is IBM's 2nd attempt. Sometimes you design it, and it just doesn't deliver. Move on!

When one looks at Digital's switch from 16 bits (PDP-11) to 32 bits (Vax 11/780), one notes that the new machines were more expensive, and about the same performance. I'd still rather have a Vax, because there are things that you can do in 32 bits that are painful in 16 (but not many).

It should be noted that throwing the address space at problems often slows it down. For example, Gosling's Emacs was ported from the Vax to the PDP-11. On the Vax, the file being edited was thrown into RAM completely. On the PDP, just a few blocks of your file were in RAM, in a paged manner. On the PDP, an insert (or delete) cause only the current page to be modified. If the current page filled up, it was split, and a new page was created. On the Vax, inserts tended to touch every page of the file - which could make the whole machine page. It was quite obviously faster on the PDP-11. No one cares about this example anymore - since machines have so much more RAM and speed. But, throwing the address space at video editing will show how bad this idea really is. Programmed I/O is smarter than having the OS do it. The program knows what it's doing, and the OS doesn't. Eventually, machines may have enough RAM and speed that no one will care, but it won't happen here at the begining of the curve.

One problem that has not been solved is the memory management unit TLB. This is the table on the chip that translated between physical and virtual memory. With 16 bits of address, 256 byte pages require only 256 entries to cover the whole address space. For 32 bit processors, the page table just doesn't fit on the chip. So, the TLB is a translation cache, and on cache miss, the OS must be called to fill it.

An alternative is to use extent lists. On my Linux system, the OS manages to keep my disk files completely contiguous 99.8% of the time. If this were done for RAM, then the number of segments that would be needed for a typical process would be small - possibly as few as four. One for text (instructions), one for initialized read only data, one for read/write data, BSS and the heap, and one for the stack. You'd need one for each DLL (shared library), but IMO, shared libraries are more trouble than they're worth, and ought to be abandoned. Removing any possibility of TLB misses would improve performance, and take much of the current mystery out of designing high performance software.

For this to work, you need the hardware vendor to produce appropriate hardware, and have at least one OS support it. The risk factor seems to have prevented this from happening so far...

Re:One of the benedits of ogg...

On: Ogg Vorbis Portables On The Way

compaines can use it in their products, without paying a royalty
This is a cost issue. I recently got a DVD player that happens to play mp3's - for $50. I can't imagine that the royalty was very much.

and without worrying about the libraries changing
I can get the source to notlame, an mp3 encoder, for free. Am I missing something?

I also recently got a radio/CD/mp3 player for my car - $145 US installed. My 15 year old car's radio died, and I decided to replace it. So now, my computer (linux), my DVD player (which is hooked up to the main stereo), and my car can play mp3 CDs.

The Ogg format seems to get better compression. But notlame seems to get about 11x compression for CDs, retaining high quality. On a 700 MB CD, I can get about 14.25 hours of music. That's at least 12, and often 14 CDs per disc.

Since I don't want to bring my $15+ CDs around in the car, I copy them anyway. mp3's allow me to use fewer disks, and have shuffle play behave like a disk jockey that knows my choices, and there are no commercials, and it even works in long tunnels...

Sure, I'd like to be able to get 17 CDs per disc, but the mp3 players are already on the market, cheap, and work.

How about a PDA?

On: Buying a Small, Light Linux Notebook Computer?

My first two laptops died after about 3 years. That's about $500/year for having a laptop. They just don't last long enough for proper amortization.

My Palm was only about $130. Yes, that means 160x160 grey. If it dies, I'm only $130 from a replacement. My screen broke, and it was only $60 to replace it.

Palm isn't Linux. But, it is a lightweight OS, and full backups of an 8 MB Palm neither take a long time, nor consume much backup media.

Unfortunately, I haven't had much luck with used laptops. Their life is short, and buying a used one robs you of some of that life.

Sharp has a Linux PDA, and IBM has released a reference platform for a PPC based PDA - for Linux. If you can't get what you want now, you should be able to get it soon.

My next recommendation is to go with a PDA with standard batteries. My Handspring uses 2 AAA batteries. I get 4 rechargables from Miejer's for $11, and a great charger for $26. If I want more spares, their cheap. The new chargers allow me to charge and forget - putting them at my convenience. If my Palm dies, I get to reuse the batteries.

Re:hubble is not obsolete yet

On: A Ground-Based Scope That Flexes For Better Focus

One can make an artificial guide star with a laser. This has already been done. And, the article talks about it.

6.5 meter main mirror, .64 secondary deformable mirror. This is a big scope. Polomar is 5 meters.

They hope to image extrasolar planets, for example, to get spectra. HST has already gotten spectra for an extrasolar planet - even though it has not imaged such an object. Step one: get the spectra for the parent star, step two: get the spectra during a transit, step three: compare.

I'm not sure why they think that a deformable secondary is better than AO afterwords.

Adaptive Optics are available for the ameteur astronomer. For example, SBIG

Re:"You wouldn't think of using a processor..."

On: Dell Dropping The Floppy

I use my 1987 Mac II - every day. A 68020 with MMU, 20 MB RAM, and a GB of disk still works for word processing, drawing, etc. The key has been upgradability. SCSI has let me add scanners, printers, disks, etc. My 5 year old PC's BIOS has an upper limit of 32 GB for hard disks. That makes it hard to find a drive that will work with it when the current drives die. The BIOS is not upgradable.

I mostly use floppies to send files to my Linux PC. Recently, I ran a memory test from a floppy. It booted DOS first. DOS is dead. It was killed by MicroSoft.

I also use my Atari 2600.

Re:evolution vs. creation

On: Racing Dinosaurs with Spoilers

Darwin's Origin Of Species predicts that intermediate species should not, in general, be found in the fossil record. This is because as a species changes, it should experience the highest environmental pressure from its closest relatives. These are the animals that are competing for the same resources.

In addition to the fossil record, one can look to living species to look for evidence. In the US, the Red Tailed Hawk shows several coloration morphs. Black ducks and Mallards can interbreed.

People seem to have problems with the idea that natural selection can ever find new solutions to problems through mutation/selection. IMO, the features that make it possible include: the power that genetics has for searching for solutions to complex problems, the length of time represented by millions of years, the multiplicative factor of having millions of individuals and the adaptability that individuals display to changes. This last allows a mutation that, by itself might be a disadvantage, to have a better chance of combining with other changes to become an advantage.

Most of evolution has to do with adaptations to new environmental conditions. The genome addresses this problem by retaining old discarded solutions to past problems. This allows relativly rapid changes.

At this point, the human animal is at risk from only one macroscopic species - other humans. That does not mean that humans will stop evolving. However, immortalitiy could change that.

Not out of numbers

On: 11 Digit Dialing Comes Home to New York

The typical blame for split area codes is put on the consumer - having more numbers. But, we're not out of numbers. The real problem is that we have competition, and every tom, dick and harry local telco, cell company, etc., gets blocks of numbers to hand out. These blocks of numbers are not a very efficient use of the number spaces.

Computers today are easily capable of dealing with the problem on a finer grained basis. For example, a cheap home PC can store and retrieve info concerning tens of millions of phone numbers in real time. Each new phone number could be allocated from a central source individually. No big deal.

Another thing that bothers me is that if you have a dial 7 area, you often can't dial 11. I should be able to dial the country code too! The phone number should be an address, not a route. I don't want to hear "You must dial a one...". If the computer knows I needed it, it should just complete the call.

On my cell phone, I always put in the dial 11, so that it still works when dialing from out of the home network. Don't dial by number, look it up by name and tell it to get a connection.

What I want is to be able to copy my phone book between my home phone, cell phone and, for editing, my computer.

Quantum efficiency

On: Improving Digital Photography

Continuous tones per pixel is good. But, one nice thing about CCDs is their high quantum efficiency. This helps in low light conditions and with fast action. As I understand it, CMOS detectors aren't as good. But, with three layers to draw on, it may be improved. Is it?

fast gravity

On: The Speed Of Gravity Revealed

One argument holds that gravity's speed must be at least 10^15th times faster than light.

Consider the Earth's orbit around the Sun. It takes light about 8 minutes to travel the distance. If gravity takes that long, then the Earth should feel the tug of gravity not where the Sun is, but where it was 8 minutes ago. This should cause gravitational drag. Though this is small over short time scales, over a 5 billion (10^9) year time scale, the Earth's orbit isn't stable.

The argument continues that since the Earth's orbit appears to be stable, gravity must move faster.

I haven't heard a really good counter argument.

If gravity really is this fast, it's going to be really hard to measure, so it's going to be really hard to use it for high speed communication.

Re:Wild ramblings...

On: The Speed Of Gravity Revealed

If an object the size of the sun suddenly acquired the 99x its mass, would it not either collapse upon itself
The sun is going the speed of light now, at least, with respect to any photon that it emits. It hasn't collapsed yet. It aquires all that mass as far as we are concerned, but it does not feel it itself, since with respect to itself, it is more or less at rest.

Re:Definitely useful

On: Secure, Efficient and Easy C programming

Most programs spend 90% of their CPU time executing 10% of their code.

I've gotten a chance to rewrite several Perl programs in C, and compared performance. On real CGI programs, the C version has outperformed Perl by a factor of 20 or 30. For longer running programs and CPU benchmarks, I'm seeing factors between 300 and 1,000. Perl is not that good for computational tasks.

One of the apps I rewrote was originally written in Perl to take advantage of Perl's hashes. The C code was 300 times faster.

So, some math. If you rewrite 10% of the app and 90% the time it takes dissappears, you are still left with the 10%. You've gained 10x of perhaps 1000x possible. There may still be 100x to go.

Taking a 1000x hit brings modern CPU performance down to 1980's Vax 11/780 performance. That level of performance was all we had, but it was mostly acceptable. It still is. My strategy has generally been to (re)write those things that need it, rather than just parts of those things. Multiple language maintenance is generally harder than any one single language maintenance.

The above may sound like Perl bashing. But, most of my code these days is written in Perl. It's most often the tool for the job.

Cheap Palms

On: Using PDAs for Dictation?

My $150 Handspring Visor Platinum has 8 MB RAM, a 33 MHz Dragonball (68000) processor with no cache or FPU. It is claimed to perform at about 5 MIPS, about what a 386/25 could do. It has a microphone, but it is connected to pins for springboard modules only. You have to have a module to use it. The placement of the microphone suggests that it's there so that a cell phone module could be built.

8 MB RAM should allow recording for over 5 minutes, uncompressed, of 22 KHz 8 bit sound. This is pretty good quality sound.

Given that you need a module for this unit anyway, you might add some hardware over just a D/A converter to make speach recognition quicker.

In any case, there's no reason that it can't be done. In fact, there are Palms with cell phones built in that can dial favorites in the address book using voice commands.

Re:on my P4 1.6GHz...

On: Electronic Life

bc
5.01*5.02-5.03/2.04*100.5+3.06+20.07-200.08+30 0.0 9/1.10
-80.80
scale=5
5.01*5.02-5.03/2.04*100.5+3.06+ 20.07-200.08+300.0 9/1.10
-126.79155

It'd be nice to get something like the right answer.

Re:I dunno

On: Electronic Life

BASIC... FOR I = 1 TO 500:NEXT I
500 empty loops took 1 second.

A modern Palm OS machine running Cbasic computes this loop about 20 times faster (10,000 loops per second). On a 1 GHz desktop, the C compiled loop

i = 1000000000;
do {
} while (--i);

executes in one second.

Byte magazine ran an April 1st article where they predicted that by 2000, PC's would be 107 MHz, switch selectable to 4.77 MHz. The joke part was the switch selectable feature, which was used to allow games that depended on the original PC's performance to continue to run. Naturally, a 486 clocked at 4.77 MHz was still much faster than an 8088 clocked at 4.77, so the switch was already an anachronism. But by 2000, a cheap PC was much faster than predicted.

An interesting question is what performance is enough? In 1981, we'd run 30+ users on a 1 MIPS machine with 4 MB RAM. 1 MIPS is enough to do word processing and surf the web, but not enough to do full screen video. A PII/350 is enough to do full screen video. Is it enough to cope with current bloatware?

Most of my cycles go to search for aliens.