> Volkswagon used to make cars that needed no liquid cooling
Hmm, this car maker sounds suspiciously similar to the one that made the Beetle, but I'm sure it's no relation. And regarding the lack of power of the Beetle engine, you could always fix that with the classic Slashdot band-aid: the Beowulf cluster. One engine in the front (I think they anticipated clustering, because they left room for one up there), one in the back, maybe one mid-ship. Should whip even their lustiest S4 Avant (380 hp) of today, right?
Windows 2000 RC2 for one, it worked inconsistently on an AMD (Compaq Presario 400 MHz notebook), sometimes BSODing at boot, other times working ok. Sure, they fixed it later, and Release is supposed to work ok.
Another was the AGP incompatibility. Many people had random lockup problems using an AGP card with an AMD. Again, maybe it got fixed by now, but it proves the point.
Finally, regarding Athlon only, the choice of mobos is simply much more limited than for Intel, and they lock you in to the chip (ok, so does the Intel).
So considering that the days of the significant price advantage of the AMD are pretty much gone, I see no reason to put up with even just the perceived potential of incompatibility. The possible speed advantage just isn't great enough to make it worthwhile, especially at the low end. And let's face it, the Celeron end of the market is still mostly where's it's at, regardless of the odd 1GHz junkies around here.
PS In the Pentium days I was definitely an AMD fan. My wife is still running a K6/200 and is quite happy with it for word processing and browsing. Back then it had both a price and performance advantage.
I think the knockers that keep harping about BSODs haven't actually used NT much lately. I have uptimes of MONTHS doing development and generally heavy use. Incidentally, I have been on W2K for about five months now (since RC2), and the same holds for it. A bloated pig, yes, but quite stable. Then again, X11 + Netscape * n isn't exactly a model of efficieny either.
The main things that bring Windows down are drivers, drivers, and drivers. Some drivers are so bad they BSOD when you look at them. I had an Oki 4W LED printer with an NT driver that BSODed at every fifth printing or so. I booted that one out the door. Once you tune up the system with good drivers, stability is very decent.
Linux certainly is no different. If there's one thing that can bring the whole thing crashing down it's immature drivers. No simple killall there, it'll coredump and all.
Overall I still prefer IE5 over Netscape. Whoever mentioned that the Netscape programmers knew what they were doing never actually had a look at the source code. It's a bloody mess, no wonder they started from SCRATCH with Mozilla. The one major thing I hate about IE though is its contrived "integration" into the OS. This means that every time IE craps out, the Windows shell craps out with the ridiculous "Explorer restarting". And of course, all the system tray icons disappear. Happens still even under W2K. That's one thing that definitely needs changing. A crappy web page or a failed transfer shouldn't bring down the GUI no matter what.
I'm the AC that posted the message to which you replied. I agree with several of your points, and the crux of the matter is basically that what I want in a PDA are:
-size of a Palm V -battery life of the Palm III (on AAA PLEASE!) -screen size and resolution of a WinCE machine (but I definitely don't need color) -processing power of a WinCE machine or better (battery life AND processing power can both be had, see StrongARM) -OS with simple and to-the-point GUI more along Palm lines, not like WinCE with its 3D shaded controls, eating away precious pixels practically (ouch, unfortunate alliteration there) mandating 320x200 resolution -I think of all the OSs I would probably prefer Epoc32, which has all the power of Win32 with much of the leannes of PalmOS -some expansion, CF is ok but a bit small for some kinds of hardware, SpringBoard approach might be preferable -any other doodaahs they can fit in without bloating the price or significantly decreasing battery life (ie sound, modem, etc)
Since this is still a pipe dream, overall I prefer the long Palm battery life along with its limitations over the kitchensink approach of WinCE with its nanoseconds of useful battery life.
Maybe the Palm form factor of the new Symbian devices will give us most of these. I've always liked the Psion platform except for the form factor--clamshell doesn't do it for me. If they can make something close to the size of the Palm V, I think that would be hard to beat.
They use a proprietary digital wireless network at 2.4 GHz, and make all kinds of use of its digital nature: intelligent call management, multiple simultaneous conversations between handsets, shared directories, etc. Maybe at some point in the future we will have a standard that works over 802.11, with handsets that automatically work with existing accesspoints and any available directories and long-distance gateways. Symbol already has a system of cordless/cell phones that works over their Spectrum24 (802.11 "compliant") wireless network, along with compatible pagers and naturally NICs. It's all pretty proprietary, though.
It would be nice if Siemens forged ahead with their lead in the home area. I have the 2415, which only allows 4 handsets (higher models allow more, it depends entirely on the base unit). They're quite affordable, I paid only $149 for a digital spread spectrum phone with digital answering machine, etc. Maybe the next iteration will have an IP address and also fetch my email and new/. postings. Would be nice.
Uwe Wolfgang Radu
Multiple textures in 3D graphics?
on
ATI Radeon 256
·
· Score: 1
Those damn arrogant third worlders, sitting at home all day surfing the web and not socializing with the rest of us. They should just get off their lazy asses and trek over to DC and read a good book for a change. It'll do them good.
But that's where the buck stops. Your attitude is dripping with contempt and disregard about a subject field I dare say you probably know less about than you might think. At least if you truly are an EE, that is. Here's why...
Unless your university was a true visionary about your eventual future employment, your exposure to the field of CS was cursory at best. You might have sat in a Fortran class, maybe even a C--or, heaven forbid--C++ class, taken some abstract assembly of an imaginary processor, and patted yourself on the back about all the computer stuff you knew. Your computer training in any case would have been heavily firmware oriented, as befits an EE. You learn to bit-twiddle, and anything that doesn't look like bit-twiddling is greeted with contempt.
Now to reality: learning the nuts and bolts of computer programming (bit-twiddling) is only the first step of a long and arduous journey. The real challenge lies in complete systems, in seeing the bigger picture, the forest as it were. EEs seldomly get past the trees. The EE equivalent would be to learn the basics about semiconductors, inductors and capacitors, and then immediately have to design an entire working system from these scraps of knowledge.
EEs are woefully unprepared to deal with large software systems. I know because I've done much of the typical EE course work, and I've worked for years with entire departments of EEs. My university required more than a cursory exposure to EE subjects for my CS major, so I've mingled with both the people and the attitudes quite freely. My previous job at TVA amongst people designing power systems (substations, hydroplants, nuclear generators) has taught me quite well the superficial exposure of EEs to computer programming.
There's a world of difference between squeezing the last cycle of performance out of an 8051, and writing objects that must interface with hundreds of other ones, and making it all actually work, and work seamlessly. The programming focus of EE and CS degrees is worlds apart, as it well should. A good CS degree will expose you to the necessary notions and ideas that hopefully in the future will get you on the right path.
Now I'm not claiming that a CS degree in itself will teach you everything you need to know. In fact, it will probably teach you very little of what you will eventually need. But it will expose you to the subject matter, and at some distant point in the future you will hopefully remember sufficiently to know where to start looking. You will have a significant leg up on the EE working with you, who will have to start from first principles on a lot of those same things. It's very analogous to the EE that might have learned about MOSFETs and op-amps at some point, but will have to do some serious recap in any real job.
Let's face it, a very significant number of EEs end up working as programmers. I don't know the exact numbers, but there must be tons (same for physicists and mathematicians, incidentally, fields I originally pursued before I "wisened up"). Most of those EEs will never again use any of the EE material they spent many nights studying. Sad maybe, but that's the economic reality. There's much higher demand for high-level software than systems design or firmware. So in that light I find it ironic and slightly pathetic that it would be the EEs making fun of CS people.
To sum up, I would hire an EE if I was working close to hardware, maybe designing a board with some micro in need of firmware, or for device driver development. I would stay as far away as possible from an EE for anything with a grander scope. In my experience EEs suck above the detail level, and CS people suck below the system level. EEs can't wrap their mind around object thinking, and CS people can't shed object thinking. These are certainly grand generalizations, but I believe they conform to the broad strokes painted by the previous posters.
Once there were the IBM CUA guidelines, and they were good. Then along came Microsoft, tried them out, and decided they could do better. As they increased their understanding of their own OS shell (Windows for the rest of you) and got a hang of all the fancy message passing, they went wild and their applications became unusable ever since. One of their worst offenders is probably Encarta, which is barely recognizable as a Windows app.
For all their faults, Apple did the best for the longest, before even they succumbed to pixmap twiddling. Their latest Quicktime player interface is a nightmare of usability. I feel this whole "real world look" where rotary knobs look and work like rotary knobs, etc, etc, etc, is the worst trend imaginable.
I'm not crying for the MDI, though. It never really conformed to the CUA, and it always felt awkward, at least to me. You'd have two windows side by side in Word, and you'd think the least it could do would be to scale them along with the parent frame when you resize the main Word window. No such luck. In the end you find yourself constantly readjusting the child windows.
One thing I think MS really got right is the Taskbar. While there might be better ways of doing it, I haven't seen one yet. Even the NeXT or Be dock on the right depend too much on icons--they don't show typical text very well, such as medium length document names, etc. However, the System Tray is getting out of hand--every app and its dog nowadays wants to put an icon in there, which then all collectively die the next time Explorer restarts.
> Although I'm not sure if I count as a real hacker, I care about themes.
You're kind of debasing your statement there. Anyway, I agree with the rest of it. I want maximum information presentation surface, so I want as few pixels as possible for control. The first thing I do after each Windows setup is to go into Appearance and reduce the size of scrollbars and menu bars. Next I move the address bar of IE into the menu bar and the links bar on the same level with the control buttons, and reduce the buttons to small icons with selective text on right. The extra space you get is unbelievable. And I'm running at 1600x1200, so it shows I'm a pixel miser.
In my experience people who like tweaking the most are the ones that just know enough to be dangerous and have figured out how to tweak. Then they think that tweaking is the essence of computing, and scoff at anybody else who doesn't tweak.
Frankly, any serious "hacker" has way too much work on his or her plate to fool around with frilly shnick-shnacks. I get into work in the morning and the first thing I do is try to figure out the bug from last night that wouldn't go away. Fiddling with skinz is the last thing on my mind. I use the Prairie Wind backdrop, which not only looks soothing, but takes up far less RAM than fancy bitmaps.
> Although I believe that the 68000 did have basic memory protection
No it didn't. The M680x0 line only received an MMU with the 68020, which came out after the OS was designed, at a time when CBM was already slacking off. Besides, they probably thought the Guru Meditation was a cute feature, why pull it for a more reliable OS?
While the Amiga OS definitely was the most advanced consumer OS of its day--multitasking, coprecessors, long file names--it also had some serious shortcomings that made progressive upgrade hard. In particular, it was fatally married to its hardware, particularly sound and graphics. Changing the graphics hardware required some serious hacks, as anybody who remembers the whole retargettable graphics (RTG) story can attest. The OS designers unfortunately never planned hardware abstraction into the software. I still have all the Amiga Kernel Manuals lying around somewhere, they're full of hardware-specific APIs. In particular the blitter was a darling of the hacker community, making possible all those cool animation demos.
So, in retrospect I must say that I absolutely loved my Amiga, I grew up programming the trusty old A500 w/o HD. But like all good things, it had to come to an end, and I had to make the painful but sensible decision to move to the PC platform. Once in a while I go back and check out the Amiga again, but nowadays I see it with different eyes. A lot like a high school sweetheart: you couldn't live without her back then, but today you kind of wonder what the excitement was all about.
Never mind that the discussion was about the Internet and not Space. But since you brought it up, let's analyze that: the US didn't know Space from its ass until it rounded up a bunch of German scientists to do its dirty work for them. The Vanguard rocket which was supposed to launch the first US satellite to catch up with the Soviets (who incidentally also profitted from the Peenemunde team), and which was originally a purely US venture, was a complete disaster. The US were a couple of decades behind the Germans in fundamental propulsion research. Lucky for them they conceded as much and took advantage of the expertise of von Braun's team, which turned out ok for the US in the end.
As regards the lazy ass Europe was sitting on while the US was putting men on the moon, it was less lazy than dirt poor from WWII. And while Germany deserves much credit for its scientific contributions of the first half of the century, it did also bring us WWII, which puts a deep shadow on those achievements. So all up, things aren't as clear cut as your American mind would like to believe.
Incidentally, while the US certainly is the flamboyant one in space operations, always showcasing its Shuttle missions--NASA's currently main claim to fame--Europe has quietly expanded its Space Agency (ESA) to the point where it's the largest satellite launcher in the world. All the while the US satellite launch program is sitting there with egg on its face.
So stop your American wet dreams right there, otherwise you run the danger of distracting from its true and genuine achievements, of which there were plenty. You simply chose the wrong ones as examples of US competence.
...he says while sitting on his potty wrapped up in his super-sized American Flag. Amerika über alles!
While you're there, ponder this: the "American Internet" was born out of paranoia, not love of freedom. Ever spent a moment considering what line of business DARPA is in? On the other hand, the Web--what most people subconsciously mean by "Internet"--was born in Europe, out of the need to share information. Not out of fear of the Soviets. Just something to think about there, Mr. Apple Pie...
For all the good parts of that movie, the constant harping on religion irritated the shit out of me. I'm sorry, but I don't remember religion being a great part of the moon landing missions, nor do I think rabbies and priests are going to be mission critical in any future Mars landings. I grew up in a very religiously conservative household, so I've seen all the pros and cons. Frankly, people that take the strict scientific view of the universe (Big Bang, Evolution, etc) and still try to bring it in harmony with the Bible make me sick. Even sicker make me religious people that try to reinterpret the Bible in scientific terms.
You take the one or the other. The Bible says the Earth was created in six days, while science has a completely different view on that, period. All that creationist science crap is just that, people that are too repressed to recognize their religion as the crutch it is, and who with each transgression of God's law relive shades of the past where their parents would smack them over the head for blasphemy or other sins. Organized religion is the most successful example of Pavlovian learning ever. Achieving obedience through the induction of guilt.
It's funny then that I walk through large companies and see Retail boxes of W2K everywhere. Lots of people aren'n even waiting for IS to install it. I guess they're all empty boxes, though, right?
I'm sorry to disappoint you, but even in this neck of the woods people all over the place are warming to W2K. It's a fact of life, it doesn't help denying it. And frankly, while we all might have other preferred OSs (I do), the vast majority of us spend the working day in front of some flavor of Windows (almost sounds comical, "flavor" of Windows). All things being equal, I'd rather it be Windows 2000. NT 4.0 wasn't bad, but its idiotic handling of hardware drove me up the walls. The PnP in W2K is the most reliable so far--again, comparatively speaking. The two main things I hate about W2K are its ravenous RAM and HD hunger. 700MB of drive space--what the heck is all in there? And that's for the Workstation, not the server. Lucky for that 17G IBM drive in there and the extra 128M of RAM I ordered.
The fact that marshalling has barely come up in the whole discussion about X shows that most people are talking out of a hole in their head, and it's not the mouth!
You simply cannot say high perfomance and RPC in the same breath, period. X is simply a specialized RPC mechanism, including all the marshalling/demarshalling of parameters that goes with it. As soon as you want to draw a line, X packages up the parameters into a packet, zips it across the network to the server, which then extracts the line parameters, figures out that you want a line, and then does the drawing. The amount of overhead involved is mind boggling. The same basic process is going on even if the server is local, except that you save the network transfer. If X were really smart, it could first check if the server is local, and if so, dispense with the entire marshalling/demarshalling process and resolve the drawing to a simple local API call. Even so, the extra steps involved in this checking would be a significant slowdown.
Consider that in comparison the original Windows drawing mechanism is much more straighforward. Drawing requests always resolve to API calls which then delegate drawing to the driver. Much more straightforward than X. Except that even that was too slow for high performance games, which led to DirectX. And even DirectX isn't the fastest thing around, according to some die-hard gamers. And yet compared to X it's orders of magnitude faster.
It would be intersting to see a breakdown of the number of machine instructions required for some given drawing primitives, first under plain X, then using memory mapped X, the plain Windows, then DirectX (maybe also NT 3.5, 4.0 and 5.0, as they've moved more of the display stuff into the kernel).
It's amusing to see hardcore Xers always defend X mainly by virtue of its remote display capabilities. And they're vocal enough that it would seem that they're the majority. But if Linux should really take off and become a popular desktop OS, remote display will hardly be used at all, by percentage of users. The great unwashed masses had a hard enough time understanding the concept of a Web page, remote data displayed locally. Getting them used to the notion of a remote program displayed locally will be a shift I wouldn't want to walk them through.
Besides, I remotely administer NT machines daily quite nicely, courtesy of a little utility called VNC. Brute force maybe, but who cares? It works well enough, doesn't bring the network to its knees (something that couldn't be said about X in the 80s), and is completely free.
I'll overlook the childish outbursts in your post and address your enthusiasm for 3dfx instead.
Since few people have seen the V5 or any other new 3dfx hardware in the flesh, I'm basing my opinions on what the guys at Tom's Hardware say. If they don't know what they're talking about, YOU certainly don't stand a chance. Go read their CeBIT report and then talk again. Besides, you don't have to be even reading any third party's opinions: check out NVidia's and 3dfx's business health and you've got it right there.
Yeah, except 3dfx is a has-been. From what I read from CeBit, their new cards suck trees. They've missed the ball, I doubt they'll ever get back in the game.
I always found it odd that Borland was such a vocal proponent of CORBA, but only included CORBA support in the Enterprise edition of Delphi. The vast majority of developers surely use Professional, but would still like to dabble with CORBA just for kicks. They seem to view CORBA mainly as a middleware technology, yet it is so much more; it can do all the mundane things COM can, and then some. Pity it's so costly to use.
I think you're trying to force a fixed point of view on the whole matter. Why would Borland have to do any conversion of any sort, either of Linux hackers to Delphi, or of Windows programmers to Linux? While some of that might indeed happen, Delphi can stand on its own legs. It is one of the most elegant development environments and languages around. And it generates commercial grade code, to boot.
Think of it this way: if Delphi for Windows had come to market a couple of years earlier, VB would have never stood a chance. Delphi offers all the ease of use and RADness of VB, while using a potent language that can scale to large projects--something VB certainly can't claim. And it creates real stand-along Windows executables; no p-code, no support DLLs etc. But since VB came first, the hordes of amateurs that jumped on it never had a reason to switch, and they went forth and multiplied. Delphi never really stood a chance.
What Borland has now is a second chance. They are going back in time to the beginnings of a new OS platform and are the first to offer a powerful RAD tool. They can then benefit from a new generation of start-up programmers that jump into Delphi because it's the only choice--like VB back then. Once they become proficient at it, they will never have a reason to switch, even if Microsoft should one day decide to port VB (ha, ha). It's all about money and leverage, not conversion. Don't forget, there are constantly new generations of programmers arising. It's just a matter of catching a big wave of them.
As an aside, I used to be a long-time C/C++ programmer (well, long-time considering my age anyway). I gave Delphi the cold shoulder for a long time--it was one of those things I was going to check out some day, just not today. One day I finally did, and I never turned back. I certainly can't claim that Object Pascal is a superior language to C++. Sometimes I miss templates, very rarely multiple inheritance, never C++'s awkward strings. It lacks the succinctness of C++, and yet I find myself going back to old code and understanding it easier than I ever did old C++ code. Delphi is a Roger Moore to C++'s Shawn Connery--less flashy, more understated, most refined. On Windows it makes working with COM a pleasure--something that can't be said about C++. In the end, I keep finding myself completing projects before my C++ brethern, especially those saddled with VC++.
> I don't think that the Kylix's C/C++ compiler > would be the optimal choice to recompile the > kernel
Why do you say that? From what I've read regarding the Borland C++ compiler ported to Linux, it beat the GNU compiler in every category when tested on an X application: compile speed (huge difference, one of Borland's strengths), executable size (almost half the size), execution speed (don't member the numbers there). Which of these would make you prefer the GNU compiler? There might be some issues regarding implementation of certain Standard C++ features, which I hear GNU is pretty strict about. Still, the Borland compiler is a very mature product, it's been worked on for many years.
Uwe Wolfgang Radu
Re:Complete lack of privacy already!
on
Database Nation
·
· Score: 2
I agree about the requirement to carry ID on your person. However, that's a different issue from the right to privacy. The obligation to be able to identify yourself does not imply the obligation to divulge any other personal information. That's where the police state begins.
Ironically--and that's a point many people don't seem to get--having a national ID system SUPPORTS strong privacy rather than undermines it. If you have a single, simple system of identification, no further information is required. On the other hand, if you don't, you must piece together all kinds of info to make you unique. It's like databases: you either have a unique key that can identify a record unambiguously, or you have composite keys consisting of several keys that together create a unique key. The latter case by its very definition leads to less privacy.
Uwe Wolfgang Radu
Complete lack of privacy already!
on
Database Nation
·
· Score: 5
The other day I went to the dentist and had to fill out one of those interminable insurance forms. Amongst other things, they wanted to know if I'm single/married/divorced/separated etc. Why the HELL do they need to know that? When are they going to start enquiring about my sexual orientation, as well as my preferred sexual positions?
This is one of my pet peeves, but Americans have no clue about personal privacy. They keep ranting against a national ID card or a national healthcare card because it would violate their privacy. Yet they think nothing of divulging their most private data to someone as inconsequential as their dentist, not to speak of using credit cards and personal checks in a system which openly laughs into their face regarding any sense of financial privacy.
Americans may rant against Europeans in any which way they like--some certainly deservedly--but regarding personal privacy they have nothing on them. While Europe is far from perfect even in the privacy issue (especially the UK), at least they try to maintain a semblance of personal privacy through the laws they pass and the way they approach the issue in general. In Germany for example, which I'm most familiar with, I can sue my dentist for breach of privacy if I feel that he is keeping data about me which he isn't entitled to. With the new digital healthcare cards I understand that I can limit the extent to which I divulge medical information even to my doctor.
Compare that to the Tennessee Department of Transportation which has included an onscure little checkbox on the driver's license renewal form, which instructs the department NOT to sell your personal information--INCLUDING YOUR MUGSHOT--to third parties. In other words, if you miss that little checkbox, which most people do, you are "authorizing" the TDOT to sell your info. If that doesn't raise your holy indignation, nothing will.
My point in all this is that we don't have to be pragmatic about privacy. There ARE things we can do to maintain and improve personal privacy, even--or rather especially--in a digital world. We have technologies that can accomplish the most amazing things: route a packet through a maze of computers from one end of the globe to another; transmit information reliably and accurately through light hours of space; write our names on the head of a pin with individual atoms; encrypt data in such a way that it would take eons to decrypt it. Yet we profess that there's nothing that can be done about the loss of privacy. It's a matter of will, not technology. We have to take the fate of our privacy out of the hands of corporations that profit from a lack of privacy, and put it into more reliable ones. Most importantly, we have to stop pretending that there's nothing we can do about it--there is, we just have to do it.
Slashdot Mirror
> Volkswagon used to make cars that needed no liquid cooling
Hmm, this car maker sounds suspiciously similar to the one that made the Beetle, but I'm sure it's no relation. And regarding the lack of power of the Beetle engine, you could always fix that with the classic Slashdot band-aid: the Beowulf cluster. One engine in the front (I think they anticipated clustering, because they left room for one up there), one in the back, maybe one mid-ship. Should whip even their lustiest S4 Avant (380 hp) of today, right?
Uwe Wolfgang Radu
Windows 2000 RC2 for one, it worked inconsistently on an AMD (Compaq Presario 400 MHz notebook), sometimes BSODing at boot, other times working ok. Sure, they fixed it later, and Release is supposed to work ok.
Another was the AGP incompatibility. Many people had random lockup problems using an AGP card with an AMD. Again, maybe it got fixed by now, but it proves the point.
Finally, regarding Athlon only, the choice of mobos is simply much more limited than for Intel, and they lock you in to the chip (ok, so does the Intel).
So considering that the days of the significant price advantage of the AMD are pretty much gone, I see no reason to put up with even just the perceived potential of incompatibility. The possible speed advantage just isn't great enough to make it worthwhile, especially at the low end. And let's face it, the Celeron end of the market is still mostly where's it's at, regardless of the odd 1GHz junkies around here.
PS In the Pentium days I was definitely an AMD fan. My wife is still running a K6/200 and is quite happy with it for word processing and browsing. Back then it had both a price and performance advantage.
Uwe Wolfgang Radu
I think the knockers that keep harping about BSODs haven't actually used NT much lately. I have uptimes of MONTHS doing development and generally heavy use. Incidentally, I have been on W2K for about five months now (since RC2), and the same holds for it. A bloated pig, yes, but quite stable. Then again, X11 + Netscape * n isn't exactly a model of efficieny either.
The main things that bring Windows down are drivers, drivers, and drivers. Some drivers are so bad they BSOD when you look at them. I had an Oki 4W LED printer with an NT driver that BSODed at every fifth printing or so. I booted that one out the door. Once you tune up the system with good drivers, stability is very decent.
Linux certainly is no different. If there's one thing that can bring the whole thing crashing down it's immature drivers. No simple killall there, it'll coredump and all.
Overall I still prefer IE5 over Netscape. Whoever mentioned that the Netscape programmers knew what they were doing never actually had a look at the source code. It's a bloody mess, no wonder they started from SCRATCH with Mozilla. The one major thing I hate about IE though is its contrived "integration" into the OS. This means that every time IE craps out, the Windows shell craps out with the ridiculous "Explorer restarting". And of course, all the system tray icons disappear. Happens still even under W2K. That's one thing that definitely needs changing. A crappy web page or a failed transfer shouldn't bring down the GUI no matter what.
Uwe Wolfgang Radu
I'm the AC that posted the message to which you replied. I agree with several of your points, and the crux of the matter is basically that what I want in a PDA are:
-size of a Palm V
-battery life of the Palm III (on AAA PLEASE!)
-screen size and resolution of a WinCE machine (but I definitely don't need color)
-processing power of a WinCE machine or better (battery life AND processing power can both be had, see StrongARM)
-OS with simple and to-the-point GUI more along Palm lines, not like WinCE with its 3D shaded controls, eating away precious pixels practically (ouch, unfortunate alliteration there) mandating 320x200 resolution
-I think of all the OSs I would probably prefer Epoc32, which has all the power of Win32 with much of the leannes of PalmOS
-some expansion, CF is ok but a bit small for some kinds of hardware, SpringBoard approach might be preferable
-any other doodaahs they can fit in without bloating the price or significantly decreasing battery life (ie sound, modem, etc)
Since this is still a pipe dream, overall I prefer the long Palm battery life along with its limitations over the kitchensink approach of WinCE with its nanoseconds of useful battery life.
Maybe the Palm form factor of the new Symbian devices will give us most of these. I've always liked the Psion platform except for the form factor--clamshell doesn't do it for me. If they can make something close to the size of the Palm V, I think that would be hard to beat.
Uwe Wolfgang Radu
They use a proprietary digital wireless network at 2.4 GHz, and make all kinds of use of its digital nature: intelligent call management, multiple simultaneous conversations between handsets, shared directories, etc. Maybe at some point in the future we will have a standard that works over 802.11, with handsets that automatically work with existing accesspoints and any available directories and long-distance gateways. Symbol already has a system of cordless/cell phones that works over their Spectrum24 (802.11 "compliant") wireless network, along with compatible pagers and naturally NICs. It's all pretty proprietary, though.
/. postings. Would be nice.
It would be nice if Siemens forged ahead with their lead in the home area. I have the 2415, which only allows 4 handsets (higher models allow more, it depends entirely on the base unit). They're quite affordable, I paid only $149 for a digital spread spectrum phone with digital answering machine, etc. Maybe the next iteration will have an IP address and also fetch my email and new
Uwe Wolfgang Radu
Get out of here, that'll never happen.
Uwe Wolfgang Radu
Those damn arrogant third worlders, sitting at home all day surfing the web and not socializing with the rest of us. They should just get off their lazy asses and trek over to DC and read a good book for a change. It'll do them good.
Uwe Wolfgang Radu
But that's where the buck stops. Your attitude is dripping with contempt and disregard about a subject field I dare say you probably know less about than you might think. At least if you truly are an EE, that is. Here's why...
Unless your university was a true visionary about your eventual future employment, your exposure to the field of CS was cursory at best. You might have sat in a Fortran class, maybe even a C--or, heaven forbid--C++ class, taken some abstract assembly of an imaginary processor, and patted yourself on the back about all the computer stuff you knew. Your computer training in any case would have been heavily firmware oriented, as befits an EE. You learn to bit-twiddle, and anything that doesn't look like bit-twiddling is greeted with contempt.
Now to reality: learning the nuts and bolts of computer programming (bit-twiddling) is only the first step of a long and arduous journey. The real challenge lies in complete systems, in seeing the bigger picture, the forest as it were. EEs seldomly get past the trees. The EE equivalent would be to learn the basics about semiconductors, inductors and capacitors, and then immediately have to design an entire working system from these scraps of knowledge.
EEs are woefully unprepared to deal with large software systems. I know because I've done much of the typical EE course work, and I've worked for years with entire departments of EEs. My university required more than a cursory exposure to EE subjects for my CS major, so I've mingled with both the people and the attitudes quite freely. My previous job at TVA amongst people designing power systems (substations, hydroplants, nuclear generators) has taught me quite well the superficial exposure of EEs to computer programming.
There's a world of difference between squeezing the last cycle of performance out of an 8051, and writing objects that must interface with hundreds of other ones, and making it all actually work, and work seamlessly. The programming focus of EE and CS degrees is worlds apart, as it well should. A good CS degree will expose you to the necessary notions and ideas that hopefully in the future will get you on the right path.
Now I'm not claiming that a CS degree in itself will teach you everything you need to know. In fact, it will probably teach you very little of what you will eventually need. But it will expose you to the subject matter, and at some distant point in the future you will hopefully remember sufficiently to know where to start looking. You will have a significant leg up on the EE working with you, who will have to start from first principles on a lot of those same things. It's very analogous to the EE that might have learned about MOSFETs and op-amps at some point, but will have to do some serious recap in any real job.
Let's face it, a very significant number of EEs end up working as programmers. I don't know the exact numbers, but there must be tons (same for physicists and mathematicians, incidentally, fields I originally pursued before I "wisened up"). Most of those EEs will never again use any of the EE material they spent many nights studying. Sad maybe, but that's the economic reality. There's much higher demand for high-level software than systems design or firmware. So in that light I find it ironic and slightly pathetic that it would be the EEs making fun of CS people.
To sum up, I would hire an EE if I was working close to hardware, maybe designing a board with some micro in need of firmware, or for device driver development. I would stay as far away as possible from an EE for anything with a grander scope. In my experience EEs suck above the detail level, and CS people suck below the system level. EEs can't wrap their mind around object thinking, and CS people can't shed object thinking. These are certainly grand generalizations, but I believe they conform to the broad strokes painted by the previous posters.
Uwe Wolfgang Radu
Once there were the IBM CUA guidelines, and they were good. Then along came Microsoft, tried them out, and decided they could do better. As they increased their understanding of their own OS shell (Windows for the rest of you) and got a hang of all the fancy message passing, they went wild and their applications became unusable ever since. One of their worst offenders is probably Encarta, which is barely recognizable as a Windows app.
For all their faults, Apple did the best for the longest, before even they succumbed to pixmap twiddling. Their latest Quicktime player interface is a nightmare of usability. I feel this whole "real world look" where rotary knobs look and work like rotary knobs, etc, etc, etc, is the worst trend imaginable.
I'm not crying for the MDI, though. It never really conformed to the CUA, and it always felt awkward, at least to me. You'd have two windows side by side in Word, and you'd think the least it could do would be to scale them along with the parent frame when you resize the main Word window. No such luck. In the end you find yourself constantly readjusting the child windows.
One thing I think MS really got right is the Taskbar. While there might be better ways of doing it, I haven't seen one yet. Even the NeXT or Be dock on the right depend too much on icons--they don't show typical text very well, such as medium length document names, etc. However, the System Tray is getting out of hand--every app and its dog nowadays wants to put an icon in there, which then all collectively die the next time Explorer restarts.
Uwe Wolfgang Radu
> Although I'm not sure if I count as a real hacker, I care about themes.
You're kind of debasing your statement there. Anyway, I agree with the rest of it. I want maximum information presentation surface, so I want as few pixels as possible for control. The first thing I do after each Windows setup is to go into Appearance and reduce the size of scrollbars and menu bars. Next I move the address bar of IE into the menu bar and the links bar on the same level with the control buttons, and reduce the buttons to small icons with selective text on right. The extra space you get is unbelievable. And I'm running at 1600x1200, so it shows I'm a pixel miser.
Uwe Wolfgang Radu
In my experience people who like tweaking the most are the ones that just know enough to be dangerous and have figured out how to tweak. Then they think that tweaking is the essence of computing, and scoff at anybody else who doesn't tweak.
Frankly, any serious "hacker" has way too much work on his or her plate to fool around with frilly shnick-shnacks. I get into work in the morning and the first thing I do is try to figure out the bug from last night that wouldn't go away. Fiddling with skinz is the last thing on my mind. I use the Prairie Wind backdrop, which not only looks soothing, but takes up far less RAM than fancy bitmaps.
Ta ta,
Uwe Wolfgang Radu
> Although I believe that the 68000 did have basic memory protection
No it didn't. The M680x0 line only received an MMU with the 68020, which came out after the OS was designed, at a time when CBM was already slacking off. Besides, they probably thought the Guru Meditation was a cute feature, why pull it for a more reliable OS?
While the Amiga OS definitely was the most advanced consumer OS of its day--multitasking, coprecessors, long file names--it also had some serious shortcomings that made progressive upgrade hard. In particular, it was fatally married to its hardware, particularly sound and graphics. Changing the graphics hardware required some serious hacks, as anybody who remembers the whole retargettable graphics (RTG) story can attest. The OS designers unfortunately never planned hardware abstraction into the software. I still have all the Amiga Kernel Manuals lying around somewhere, they're full of hardware-specific APIs. In particular the blitter was a darling of the hacker community, making possible all those cool animation demos.
So, in retrospect I must say that I absolutely loved my Amiga, I grew up programming the trusty old A500 w/o HD. But like all good things, it had to come to an end, and I had to make the painful but sensible decision to move to the PC platform. Once in a while I go back and check out the Amiga again, but nowadays I see it with different eyes. A lot like a high school sweetheart: you couldn't live without her back then, but today you kind of wonder what the excitement was all about.
Uwe Wolfgang Radu
Never mind that the discussion was about the Internet and not Space. But since you brought it up, let's analyze that: the US didn't know Space from its ass until it rounded up a bunch of German scientists to do its dirty work for them. The Vanguard rocket which was supposed to launch the first US satellite to catch up with the Soviets (who incidentally also profitted from the Peenemunde team), and which was originally a purely US venture, was a complete disaster. The US were a couple of decades behind the Germans in fundamental propulsion research. Lucky for them they conceded as much and took advantage of the expertise of von Braun's team, which turned out ok for the US in the end.
As regards the lazy ass Europe was sitting on while the US was putting men on the moon, it was less lazy than dirt poor from WWII. And while Germany deserves much credit for its scientific contributions of the first half of the century, it did also bring us WWII, which puts a deep shadow on those achievements. So all up, things aren't as clear cut as your American mind would like to believe.
Incidentally, while the US certainly is the flamboyant one in space operations, always showcasing its Shuttle missions--NASA's currently main claim to fame--Europe has quietly expanded its Space Agency (ESA) to the point where it's the largest satellite launcher in the world. All the while the US satellite launch program is sitting there with egg on its face.
So stop your American wet dreams right there, otherwise you run the danger of distracting from its true and genuine achievements, of which there were plenty. You simply chose the wrong ones as examples of US competence.
Uwe Wolfgang Radu
...he says while sitting on his potty wrapped up in his super-sized American Flag. Amerika über alles!
While you're there, ponder this: the "American Internet" was born out of paranoia, not love of freedom. Ever spent a moment considering what line of business DARPA is in? On the other hand, the Web--what most people subconsciously mean by "Internet"--was born in Europe, out of the need to share information. Not out of fear of the Soviets. Just something to think about there, Mr. Apple Pie...
Uwe Wolfgang Radu
For all the good parts of that movie, the constant harping on religion irritated the shit out of me. I'm sorry, but I don't remember religion being a great part of the moon landing missions, nor do I think rabbies and priests are going to be mission critical in any future Mars landings. I grew up in a very religiously conservative household, so I've seen all the pros and cons. Frankly, people that take the strict scientific view of the universe (Big Bang, Evolution, etc) and still try to bring it in harmony with the Bible make me sick. Even sicker make me religious people that try to reinterpret the Bible in scientific terms.
You take the one or the other. The Bible says the Earth was created in six days, while science has a completely different view on that, period. All that creationist science crap is just that, people that are too repressed to recognize their religion as the crutch it is, and who with each transgression of God's law relive shades of the past where their parents would smack them over the head for blasphemy or other sins. Organized religion is the most successful example of Pavlovian learning ever. Achieving obedience through the induction of guilt.
Uwe Wolfgang Radu
It's funny then that I walk through large companies and see Retail boxes of W2K everywhere. Lots of people aren'n even waiting for IS to install it. I guess they're all empty boxes, though, right?
Uwe Wolfgang Radu
I'm sorry to disappoint you, but even in this neck of the woods people all over the place are warming to W2K. It's a fact of life, it doesn't help denying it. And frankly, while we all might have other preferred OSs (I do), the vast majority of us spend the working day in front of some flavor of Windows (almost sounds comical, "flavor" of Windows). All things being equal, I'd rather it be Windows 2000. NT 4.0 wasn't bad, but its idiotic handling of hardware drove me up the walls. The PnP in W2K is the most reliable so far--again, comparatively speaking. The two main things I hate about W2K are its ravenous RAM and HD hunger. 700MB of drive space--what the heck is all in there? And that's for the Workstation, not the server. Lucky for that 17G IBM drive in there and the extra 128M of RAM I ordered.
Uwe Wolfgang Radu
The fact that marshalling has barely come up in the whole discussion about X shows that most people are talking out of a hole in their head, and it's not the mouth!
You simply cannot say high perfomance and RPC in the same breath, period. X is simply a specialized RPC mechanism, including all the marshalling/demarshalling of parameters that goes with it. As soon as you want to draw a line, X packages up the parameters into a packet, zips it across the network to the server, which then extracts the line parameters, figures out that you want a line, and then does the drawing. The amount of overhead involved is mind boggling. The same basic process is going on even if the server is local, except that you save the network transfer. If X were really smart, it could first check if the server is local, and if so, dispense with the entire marshalling/demarshalling process and resolve the drawing to a simple local API call. Even so, the extra steps involved in this checking would be a significant slowdown.
Consider that in comparison the original Windows drawing mechanism is much more straighforward. Drawing requests always resolve to API calls which then delegate drawing to the driver. Much more straightforward than X. Except that even that was too slow for high performance games, which led to DirectX. And even DirectX isn't the fastest thing around, according to some die-hard gamers. And yet compared to X it's orders of magnitude faster.
It would be intersting to see a breakdown of the number of machine instructions required for some given drawing primitives, first under plain X, then using memory mapped X, the plain Windows, then DirectX (maybe also NT 3.5, 4.0 and 5.0, as they've moved more of the display stuff into the kernel).
It's amusing to see hardcore Xers always defend X mainly by virtue of its remote display capabilities. And they're vocal enough that it would seem that they're the majority. But if Linux should really take off and become a popular desktop OS, remote display will hardly be used at all, by percentage of users. The great unwashed masses had a hard enough time understanding the concept of a Web page, remote data displayed locally. Getting them used to the notion of a remote program displayed locally will be a shift I wouldn't want to walk them through.
Besides, I remotely administer NT machines daily quite nicely, courtesy of a little utility called VNC. Brute force maybe, but who cares? It works well enough, doesn't bring the network to its knees (something that couldn't be said about X in the 80s), and is completely free.
Uwe Wolfgang Radu
I'll overlook the childish outbursts in your post and address your enthusiasm for 3dfx instead.
Since few people have seen the V5 or any other new 3dfx hardware in the flesh, I'm basing my opinions on what the guys at Tom's Hardware say. If they don't know what they're talking about, YOU certainly don't stand a chance. Go read their CeBIT report and then talk again. Besides, you don't have to be even reading any third party's opinions: check out NVidia's and 3dfx's business health and you've got it right there.
Uwe Wolfgang Radu
Yeah, except 3dfx is a has-been. From what I read from CeBit, their new cards suck trees. They've missed the ball, I doubt they'll ever get back in the game.
Uwe Wolfgang Radu
I always found it odd that Borland was such a vocal proponent of CORBA, but only included CORBA support in the Enterprise edition of Delphi. The vast majority of developers surely use Professional, but would still like to dabble with CORBA just for kicks. They seem to view CORBA mainly as a middleware technology, yet it is so much more; it can do all the mundane things COM can, and then some. Pity it's so costly to use.
Uwe Wolfgang Radu
I think you're trying to force a fixed point of view on the whole matter. Why would Borland have to do any conversion of any sort, either of Linux hackers to Delphi, or of Windows programmers to Linux? While some of that might indeed happen, Delphi can stand on its own legs. It is one of the most elegant development environments and languages around. And it generates commercial grade code, to boot.
Think of it this way: if Delphi for Windows had come to market a couple of years earlier, VB would have never stood a chance. Delphi offers all the ease of use and RADness of VB, while using a potent language that can scale to large projects--something VB certainly can't claim. And it creates real stand-along Windows executables; no p-code, no support DLLs etc. But since VB came first, the hordes of amateurs that jumped on it never had a reason to switch, and they went forth and multiplied. Delphi never really stood a chance.
What Borland has now is a second chance. They are going back in time to the beginnings of a new OS platform and are the first to offer a powerful RAD tool. They can then benefit from a new generation of start-up programmers that jump into Delphi because it's the only choice--like VB back then. Once they become proficient at it, they will never have a reason to switch, even if Microsoft should one day decide to port VB (ha, ha). It's all about money and leverage, not conversion. Don't forget, there are constantly new generations of programmers arising. It's just a matter of catching a big wave of them.
As an aside, I used to be a long-time C/C++ programmer (well, long-time considering my age anyway). I gave Delphi the cold shoulder for a long time--it was one of those things I was going to check out some day, just not today. One day I finally did, and I never turned back. I certainly can't claim that Object Pascal is a superior language to C++. Sometimes I miss templates, very rarely multiple inheritance, never C++'s awkward strings. It lacks the succinctness of C++, and yet I find myself going back to old code and understanding it easier than I ever did old C++ code. Delphi is a Roger Moore to C++'s Shawn Connery--less flashy, more understated, most refined. On Windows it makes working with COM a pleasure--something that can't be said about C++. In the end, I keep finding myself completing projects before my C++ brethern, especially those saddled with VC++.
Uwe Wolfgang Radu
> I don't think that the Kylix's C/C++ compiler
> would be the optimal choice to recompile the
> kernel
Why do you say that? From what I've read regarding the Borland C++ compiler ported to Linux, it beat the GNU compiler in every category when tested on an X application: compile speed (huge difference, one of Borland's strengths), executable size (almost half the size), execution speed (don't member the numbers there). Which of these would make you prefer the GNU compiler? There might be some issues regarding implementation of certain Standard C++ features, which I hear GNU is pretty strict about. Still, the Borland compiler is a very mature product, it's been worked on for many years.
Uwe Wolfgang Radu
I agree about the requirement to carry ID on your person. However, that's a different issue from the right to privacy. The obligation to be able to identify yourself does not imply the obligation to divulge any other personal information. That's where the police state begins.
Ironically--and that's a point many people don't seem to get--having a national ID system SUPPORTS strong privacy rather than undermines it. If you have a single, simple system of identification, no further information is required. On the other hand, if you don't, you must piece together all kinds of info to make you unique. It's like databases: you either have a unique key that can identify a record unambiguously, or you have composite keys consisting of several keys that together create a unique key. The latter case by its very definition leads to less privacy.
Uwe Wolfgang Radu
The other day I went to the dentist and had to fill out one of those interminable insurance forms. Amongst other things, they wanted to know if I'm single/married/divorced/separated etc. Why the HELL do they need to know that? When are they going to start enquiring about my sexual orientation, as well as my preferred sexual positions?
This is one of my pet peeves, but Americans have no clue about personal privacy. They keep ranting against a national ID card or a national healthcare card because it would violate their privacy. Yet they think nothing of divulging their most private data to someone as inconsequential as their dentist, not to speak of using credit cards and personal checks in a system which openly laughs into their face regarding any sense of financial privacy.
Americans may rant against Europeans in any which way they like--some certainly deservedly--but regarding personal privacy they have nothing on them. While Europe is far from perfect even in the privacy issue (especially the UK), at least they try to maintain a semblance of personal privacy through the laws they pass and the way they approach the issue in general. In Germany for example, which I'm most familiar with, I can sue my dentist for breach of privacy if I feel that he is keeping data about me which he isn't entitled to. With the new digital healthcare cards I understand that I can limit the extent to which I divulge medical information even to my doctor.
Compare that to the Tennessee Department of Transportation which has included an onscure little checkbox on the driver's license renewal form, which instructs the department NOT to sell your personal information--INCLUDING YOUR MUGSHOT--to third parties. In other words, if you miss that little checkbox, which most people do, you are "authorizing" the TDOT to sell your info. If that doesn't raise your holy indignation, nothing will.
My point in all this is that we don't have to be pragmatic about privacy. There ARE things we can do to maintain and improve personal privacy, even--or rather especially--in a digital world. We have technologies that can accomplish the most amazing things: route a packet through a maze of computers from one end of the globe to another; transmit information reliably and accurately through light hours of space; write our names on the head of a pin with individual atoms; encrypt data in such a way that it would take eons to decrypt it. Yet we profess that there's nothing that can be done about the loss of privacy. It's a matter of will, not technology. We have to take the fate of our privacy out of the hands of corporations that profit from a lack of privacy, and put it into more reliable ones. Most importantly, we have to stop pretending that there's nothing we can do about it--there is, we just have to do it.
Uwe Wolfgang Radu