Herb Sutter wrote an article that describes this much better than I can, but basically, you're not far off the mark...you shouldn't be seeing or expecting any major increases in performance via a faster clock, but by more cores and more multithreading.
It wouldn't surprise me a bit if Apple releases new machines to totally obscure the actual ghz number for some other selling point like number of cores or somesuch.
I wrote a chart control in C++ where I used bitshifting instead of multiply/divide and that is the only time (so far) where I was trying to squeeze maximum performance out of my app.
For the most part, unless you're writing anything with hard deadlines (RTOS', graphical stuff where visual matters, etc.) the time to look at the assembly is time wasted; better to get your algorithms right than to cut 10 instructions to two. Besides, most of the time, especially if you're doing "business"-related work, you're going to probably be bound by the disk or the network or other factors that makes the point moot.
So while good assembly is important, I'll trust the compiler does a good enough job for my purposes. However, that's not to say you can't help it along by passing by reference, using the STL properly, etc. Thus I try to strike a balance between readable/maintainable code, and performance. However, given that most of the things I write are business apps, maintenability will always always win out.
So this is quite interesting; I'll be very interested to see their implementation. With IBM's MQ, their big thing was that there was something like 8 function calls, tops. I can still see the IBM guy talking about how simple it was, listing out the functions.
What he neglected to mention was that each of those function calls takes *structures* as arguments...some of those structures had a dozen+ members.
To be fair, MQ is pretty slick; I wrote a connector between the webserver and an ES/9000 mainframe using MQ. No EBCIDIC tranlation, no worrying about going from TCP/IP to SNA...all was handled by MQ.
The other big thing (and I'm sure this is where all this work is coming from), is that, on a Windows platform, I think MQ cost around $5000 (circa 1999). On the mainframe, they charge *by the cycle*. I don't remember the exact figures, but the head of the mainframe group said that it cost something in the high five figures just to get the cartridge, and then there was both a monthly support cost, and then the per-cycle cost.
When testing my program for load, I would send a couple of big books I had gotten from the Gutenberg project on a round trip from the NT box to the mainframe, and do it all weekend. My manager was presented with a "bill" for around $10,000 in MQ time on the mainframe. Yowza.
And in addition to all that, there were the Candle consultants who helped set the topography up. For an end-to-end system, it sure had a lot of people involved, and a lot of moving pieces.
The upshot was that it did what it was supposed to do, and we handled millions of requests through it every single day for years. Still one of the pieces of code I'm most proud of.
I provided support for a guy's small business about 20 miles away from my house; it would take at least 30 minutes to get there and help him keep his POS machines running Win95 alive for another year. I did everything from installing a router to helping him set up FrontPage, to cleaning junk out of a mouse. Typically, he'd take me out to dinner and pay me about $50 (no, never really had a set price, which is my mistake).
The proverbial straw came when he asked if I could put a machine that had never been on the network on it, so he wouldn't have to use the modem to connect to everything (while every other machine was running off a shared dsl line). I opened the box and discovered that it was a pci/ISA machine, and the network card he had was an ISA one.
To make a long story short, I took it as a challenge, and spent the rest of the night trying to remember irqs/offsets, and all those bizzare things I used to do back in '94/95. He was a little concerned, thinking that messing with jumpers meant you were doing something wrong. No, I assured him, getting stuff to work on this vintage of machine required this kind of stuff.
Four hours of work, all old school setup, and I got it working. There was also an element of danger, as this machine had all the quickbook stuff, so when it booted into Win95 and then froze, I was more than a little nervous. But, yes, I got it to work.
I got $50.
I didn't actually see the check until I was in the car, and then I screamed, went home, called a mutual friend, screamed some more about what a cheap bastard the guy was, and somehow this all got back to the guy. Hasn't called me since.
So while it's my fault for simply not stating up front a number, when confronted with the "real" value of what I did, he simply shrugged and decided it wasn't worth it to call anymore. Dunno if his machines are still running. Don't care.
Still, it was an easy $50 or so bucks most of the time; was nice to have that tiny extra bit of $.
More importantly, will it have a model wearing a plaid miniskirt and beehive hairdo smiling, extolling how she can now service multiple jobs at the same time?
Think about when the iPod Shuffle was announced...Apple had their website ready to go with quicktime vr movies, product shots, as well as "action" shots of real people with the device. Clearly, a *lot* of people knew about it by that time because it was a. already built so that b. it could be photographed (even if it were a non-working unit, they wouldn't take pictures of something that would change drastically before launch).
So if they're really cranking out the G5 PBs now, then there's gotta be a bunch of photographers and models running around offices and parks, probably in California, posing for those action shots.
Now the trick is to find them. I wonder when they're getting the people to pose, if they ask them if they're Mac users/fanatics, so that they don't get someone who, 5 seconds after wrapping up the shoot, goes online and says "OMG! You will *NOT* believe what I was just using an hour ago!!!!!!!"
While the last promotion was going on, deep in the winter, I would walk into the store and they'd still be selling Pepsi with variations of "Summer fun!" on them. Maybe they had the Apple/Pepsi bottles in the back, but I guess they just have them in a queue; sell the old ones first.
Can PepsiCo "demand" or "insist" that these bottles be put up first, or are they really at the mercy of the owner once the bottles hit the store?
Yesterday it was an iPod Shuffle being pulled apart, today it's the Mac Mini being disemboweled. While I definately RTFA and enjoyed TFA for both, where did they get these units so early? The Apple store in town says they won't get any till the beginning of next month and I haven't heard of anyone getting either from Apple's online store yet.
As it was unlawfully taken away from Panix, can't they take the same or similar steps and simply take their domain back? Presumably the jerks who did this in the first place aren't exactly in a position to cry foul.
As an aside, if not Panix, then why not IBM or ebay or Amazon? Is it a case of "nobody's tried this yet" or do the "big names" have something that everybody else doesn't? I would think microsoft.com is presumably just as vunerable to domain theft as joespixoftoilets.com. Also, while if the folks who did this to Panix were Aussies, and if they tried it with IBM or Microsoft, those companies would have lawyers literally pounding on the door of the offending registrar within minutes, who's to say the person can't do it at some ISP in Russia or some other country where the likelyhood of having any legal weight is practically nil?
I worked at a bank in the early 90s that had a trading system based on SQL Server and the client was written in Visual Basic 3. Apart from every other bad design choice in this system (I inherited it when the designers got promoted and started working on another, even bigger system), the all important record counter was an integer, so when trade 32768 was posted, the application crashed, and simply could not be started again, because the first thing it did was try to show the current total (it was written for operators to use, not traders). Worse was that the counter variable wasn't a global, and it was often times a stack variable, and always with a different name (sometimes iCounter, sometimes iCount, sometimes x).
The upshot was that I was able to convince management to totally scrap it and allow me to write a new one. The downside was that the idiot who designed the original system went on to spend 100 million dollars on this new, grandious system that too was eventually scrapped, but he knew long before that his turkey wasn't going to fly, so he quit and became a lead architect at some other company.
IBM sells a product simply called VM. Actually, I guess it's more of a lease option, as it's only available for mainframes (and I used it on an ES/9000...one of the biggest mainframe (read: MVS/CICS) machines around). It's cool in that you could assign separate processors separate copies of the OS, unlike VMware which has a "host" operating system and then various Guests. There's still a bit of low-level software, but for us it was seemless (which, given how much everything associated with this machine cost, had better have been).
Interestingly, this brought to mind the Pink operating system that IBM and Apple were working on way-back-when(tm). The idea, if I remember correctly, was to have a low level OS kernel that could run multiple personalities...they talked about a MacOS personality (back when System 8 was still being developed), OS/2 and probably some flavor of Unix.
I remember being at what I believe was the last Unix Convention at the Javits Center in NYC around '92 or '93 and they (IBM) had a prototype Power box that purported to be running a super super early pre-alpha version of it. The guy standing by it wouldn't let me touch it, and all he said he could do was run a "DIR" on what was supposed to be the OS/2 personality (no Mac one in sight, for the obvious reason there never was one). He also mentioned that there was a second box, but they couldn't get it to boot.
*Sigh*... strange times. Full of promises yet to be fulfilled. But as someone else pointed out, now that OS X is essentialy Unix, there would be precious little reason to go back to the "personality" scheme. I rather think they'd bring out some kick-ass server type box running multiple copies of OS X server, if that is in fact what they're trying to do.
I was actually under the impression it was just going to be a dual core PPC, but I RTFA off os OSNews.com a couple of days ago and I don't really remember it.
I have a BFA in film from a school that has a couple of entries in the Jargon file, but that's about it. I am totally self-taught in terms of programming, and I've now been a Windows/Unix C++ programmer for seven years.
I started by writing Excel 4 macros, moved up to VB, then went to C, later C++. I also learned a lot of SQL, where I can now write sprocs in T-SQL for Sybase/SQL Server, and PL/SQL for Oracle.
In every "move up", it was seeing a need for the next level up when the stuff at the previous level wasn't good enough (needed custom VBXs for VB, etc.) It wasn't lying, we really did need this stuff, and I was willing to step up to the plate to do it. I put in a *lot* of long hours. I read and read and read and played every single day. In all of 1996, I took July 4th, Thanksgiving, and Christmas off...I worked all the others.
I was also upfront and made mistakes (especially when threads got involved), but I worked like hell to fix, and more importantly, *totally* understand the actual issues. I think it also helped that we didn't have things like the STL to use, so I had to roll my own binary trees and linked lists by hand...a great way to learn how to use pointers.
I'm not so sure this method will work for anyone else, as everyone has their own journey to take. But I worked with a lot of others who had various art degrees who were also in IT, and they did it basically the same way as me...lots of grunt work to understand what was going on, and could demonstrate it at both the interview and on the job.
So I would say that it doesn't matter. I also will say that as one who has interviewed a lot of people, I never take into consideration the "Education" section of the resume. You either know the answer to my questions, or can at least try to reason a good guess, or you don't. You can have a PhD in CompSci from Harvard, but if you can't tell me how to update a table in the database, then it's not going to help.
While it's true that Microsoft will always go for lock in, and their own language makes that easier (C#), I hate to admit it, but their 7.1 C++ compiler is quite good, and compiles every wacky template I throw at it. It's supposedly been able to compile Andresceu's Loki library, as well as Boost++, both very taxing on compilers. I think this is more a case of Herb Sutter's championing than anything else. So while you're probably right, at least their C++ compiler does kick ass.
For the best example of outspoken developers keeping a product alive, look no further than FoxPro. MS bought Foxbase to get their Rushmore technology (which I believe was bitmapped indexes brought to the.DBF level) to put into Access 2.0, and then promply announced that FoxPro was dead. The FoxPro community howled long and hard and MS has basically had to keep it alive ever since, always as a separate product, but it's gotten a lot of tweaks and upgrades to make it look like other MS products (read: tabs).
This is another product you won't hear pass anyone's lips from MS, but it wouldn't surprise me if we see a FoxPro 9 somewhere +/- the time VS2005 is released.
Yeah, my mistake...it was a PS/2 Model 95 with a 60mhz Pentium chip. Heh...that thing ran Netware 3.12 and serviced both IPX and TCP/IP for a couple hundred people at my school as well as being both a mac/pc file server and handling I think at least 3 printers. It probably ran extra hot from being at 100% cpu utilization all the time.
Of course, the original PPro's (60/66mhz) were hotter than a 40 watt light bulb. I made the mistake of touching just the side of one when working on a PS/2 that had one. That chip was the G5 of its time...how are they *ever* going to make that thing into a portable?
Heh...mentioning MCA reminded me that there was a board available for Model 95s that turned the machine into an extremely small mainframe, but could still run all MVS apps (including CICS and DB/2). Dunno about the IO on the board, but I guess it was IBMs acknowledgement that it could be done (I believe it was a "proof of concept" that never actually made it into the channels).
As far as IO, I worked at a bank that had an ES/9000...one of the biggest around. Thing was half the entire floor of the building, in all its water cooled glory. The way it computed your balance was to go through the VSAM files, find you, then start from the very first entry and add/subtract each payment/debit until it reached the end, and you had your balance. Sure it would cache it for awhile (I think a couple of days), but eventually it'd do the whole add/subtract thing all over again.
I could run execute the cobol program from the web, and then get the result back via MQSeries. Never saw it take more than a second...when I was at a dos console, I'd hit enter and it'd bang back the answer without delay. Fastest thing I'd ever seen.
When I went to another job where they were facing a similar problem (having to find/compute records in massive files) I suggested that a mainframe was the way to go. Being a.com, they didn't even consider it (of course, it's not like anybody can afford an ES/9000...though I suppose the zSeries is pretty good), and as such they never had a solution that gave acceptable performance.
No, you're right, and I thought about qualifying that in the post. I would not want to try (actually, maybe I would like to try) to process GE's payroll on an emulated IBM 1401 or 360.
I think it was more the fact that it *can* be run; that the ability to run CICS, once only available on a 3090 or whatever, could also run on a PC running a mainframe emulator is what I was thinking.
But you're right, the only thing that performs as good as a mainframe is another mainframe.
I googled RETAIN and from the couple of sites that I read, it jogged the memory and I realize that I think I'd been on the receiving end of it...I worked at a government agency which had an AS/400 back in '91 and one day our local IBM rep shows up with a new 9 track drive and says he's going to install it. I said I don't think anyone had mentioned it, and he replied that the AS/400 had detected a problem and sent a message to IBM. Now I knew why there was that extra phone line attached it it.
I think RETAIN was/is behind those cool radio-based units IBM reps carried...they had a 2 line LCD screen, a chicklet-style qwerty keyboard, but it was like an early Blackberry...the rep could find out where he had to go next, what the problem was, what parts were necessary, etc. I thought it was *very* *very* cool back then (actually, it's still cool), when I was still struggling to get a 9600bps modem working with Procomm.
Because these books talk about a time when punch cards were still all the rage, and because my Ask Slashdot article was rejected, I'll ask here:
Does anyone still use punch cards? I know some states used punch cards for the Nov 2 election, but I'm wondering if there are still decks of cards at companies waiting to be run through and the output printed on green-n-white paper.
It's not a criticism or putdown question, I can believe there are some jobs on some equipment that just can't (or won't) be ported to something newer, and "what worked for us back then works for us now."
I don't think I ever read this book (born in 1970), but flipping through the pages, it makes me realize what computers still mean to my folks; batch cards, mag tapes, green-n-white printouts.
Therein lies the rub; to my folks, any computer that can be fit in a single box and doesn't live in a raised-floor room, is a toy. It's actually very black and white for them..."yes it's all very nice what those toys can do for the movies, but it takes a *computer* to process GE's payroll."
It also reminds me of when a friend of mine brought his dad in to work to show him what he did. His dad was a serious old school programmer for custom chips for Navy jets. He looked it too...checkered shirt, crew cut, pocket protector (first time I'd ever seen one). My friend shows him the *cough* Powerbuilder app we'd be working on, with its buttons and datawindows, etc., and his dad just went *pft* and waved his hand.
The fact that I can run emulators of any of those systems and they run 10x faster has never made a dent in my folks opinion. As far as they can see, and as far as my friend's dad can see, we're just playing with toys.
True enough, Renderman is the standard as far as I know. But because Pixar is both producer and consumer of said software (they also make some other tools that I'm forgetting the names of at the moment), they also have access to the developers, so if the Nemo team wanted water to look "just so", they'd merely have to walk over to the appropriate person's office, and maybe if it wasn't in there before, hey, a little coding and there you go. Even a paying customer like Disney wouldn't get that kind of service.
I have an original clamshell iBook with 64meg of ram...far too small in terms of resources, I found, for OS X. Linux (specifically YDL 3) ran acceptably on it and am able to do a lot more with it than what I could do under OS 9 (ironically, a lot of stuff that winds up running on OS X). Gnome doesn't run *great* on it, but it's acceptable. Plus with YDL it was easy to configure the Airport card in it, so I can sit anywhere and use it.
I think you'll find most people run Linux on older macs to revive and get some more use out of older hardware...I wish I hadn't thrown out my 68k-based Mac so I could try NetBSD on it.
I could also see where someone might have a very specific need to run in 64-bit *now*, instead of waiting for Tiger next year; then a G5-specific Linux kernel like YDL 4 would fit the bill.
Slashdot Mirror
Herb Sutter wrote an article that describes this much better than I can, but basically, you're not far off the mark...you shouldn't be seeing or expecting any major increases in performance via a faster clock, but by more cores and more multithreading.
It wouldn't surprise me a bit if Apple releases new machines to totally obscure the actual ghz number for some other selling point like number of cores or somesuch.
I wrote a chart control in C++ where I used bitshifting instead of multiply/divide and that is the only time (so far) where I was trying to squeeze maximum performance out of my app.
For the most part, unless you're writing anything with hard deadlines (RTOS', graphical stuff where visual matters, etc.) the time to look at the assembly is time wasted; better to get your algorithms right than to cut 10 instructions to two. Besides, most of the time, especially if you're doing "business"-related work, you're going to probably be bound by the disk or the network or other factors that makes the point moot.
So while good assembly is important, I'll trust the compiler does a good enough job for my purposes. However, that's not to say you can't help it along by passing by reference, using the STL properly, etc. Thus I try to strike a balance between readable/maintainable code, and performance. However, given that most of the things I write are business apps, maintenability will always always win out.
So this is quite interesting; I'll be very interested to see their implementation. With IBM's MQ, their big thing was that there was something like 8 function calls, tops. I can still see the IBM guy talking about how simple it was, listing out the functions.
What he neglected to mention was that each of those function calls takes *structures* as arguments...some of those structures had a dozen+ members.
To be fair, MQ is pretty slick; I wrote a connector between the webserver and an ES/9000 mainframe using MQ. No EBCIDIC tranlation, no worrying about going from TCP/IP to SNA...all was handled by MQ.
The other big thing (and I'm sure this is where all this work is coming from), is that, on a Windows platform, I think MQ cost around $5000 (circa 1999). On the mainframe, they charge *by the cycle*. I don't remember the exact figures, but the head of the mainframe group said that it cost something in the high five figures just to get the cartridge, and then there was both a monthly support cost, and then the per-cycle cost.
When testing my program for load, I would send a couple of big books I had gotten from the Gutenberg project on a round trip from the NT box to the mainframe, and do it all weekend. My manager was presented with a "bill" for around $10,000 in MQ time on the mainframe. Yowza.
And in addition to all that, there were the Candle consultants who helped set the topography up. For an end-to-end system, it sure had a lot of people involved, and a lot of moving pieces.
The upshot was that it did what it was supposed to do, and we handled millions of requests through it every single day for years. Still one of the pieces of code I'm most proud of.
I provided support for a guy's small business about 20 miles away from my house; it would take at least 30 minutes to get there and help him keep his POS machines running Win95 alive for another year. I did everything from installing a router to helping him set up FrontPage, to cleaning junk out of a mouse. Typically, he'd take me out to dinner and pay me about $50 (no, never really had a set price, which is my mistake).
The proverbial straw came when he asked if I could put a machine that had never been on the network on it, so he wouldn't have to use the modem to connect to everything (while every other machine was running off a shared dsl line). I opened the box and discovered that it was a pci/ISA machine, and the network card he had was an ISA one.
To make a long story short, I took it as a challenge, and spent the rest of the night trying to remember irqs/offsets, and all those bizzare things I used to do back in '94/95. He was a little concerned, thinking that messing with jumpers meant you were doing something wrong. No, I assured him, getting stuff to work on this vintage of machine required this kind of stuff.
Four hours of work, all old school setup, and I got it working. There was also an element of danger, as this machine had all the quickbook stuff, so when it booted into Win95 and then froze, I was more than a little nervous. But, yes, I got it to work.
I got $50.
I didn't actually see the check until I was in the car, and then I screamed, went home, called a mutual friend, screamed some more about what a cheap bastard the guy was, and somehow this all got back to the guy. Hasn't called me since.
So while it's my fault for simply not stating up front a number, when confronted with the "real" value of what I did, he simply shrugged and decided it wasn't worth it to call anymore. Dunno if his machines are still running. Don't care.
Still, it was an easy $50 or so bucks most of the time; was nice to have that tiny extra bit of $.
More importantly, will it have a model wearing a plaid miniskirt and beehive hairdo smiling, extolling how she can now service multiple jobs at the same time?
Think about when the iPod Shuffle was announced...Apple had their website ready to go with quicktime vr movies, product shots, as well as "action" shots of real people with the device. Clearly, a *lot* of people knew about it by that time because it was a. already built so that b. it could be photographed (even if it were a non-working unit, they wouldn't take pictures of something that would change drastically before launch).
So if they're really cranking out the G5 PBs now, then there's gotta be a bunch of photographers and models running around offices and parks, probably in California, posing for those action shots.
Now the trick is to find them. I wonder when they're getting the people to pose, if they ask them if they're Mac users/fanatics, so that they don't get someone who, 5 seconds after wrapping up the shoot, goes online and says "OMG! You will *NOT* believe what I was just using an hour ago!!!!!!!"
While the last promotion was going on, deep in the winter, I would walk into the store and they'd still be selling Pepsi with variations of "Summer fun!" on them. Maybe they had the Apple/Pepsi bottles in the back, but I guess they just have them in a queue; sell the old ones first.
Can PepsiCo "demand" or "insist" that these bottles be put up first, or are they really at the mercy of the owner once the bottles hit the store?
"Death to others"
Yesterday it was an iPod Shuffle being pulled apart, today it's the Mac Mini being disemboweled. While I definately RTFA and enjoyed TFA for both, where did they get these units so early? The Apple store in town says they won't get any till the beginning of next month and I haven't heard of anyone getting either from Apple's online store yet.
Maybe they're the leftovers from the promos?
As it was unlawfully taken away from Panix, can't they take the same or similar steps and simply take their domain back? Presumably the jerks who did this in the first place aren't exactly in a position to cry foul.
As an aside, if not Panix, then why not IBM or ebay or Amazon? Is it a case of "nobody's tried this yet" or do the "big names" have something that everybody else doesn't? I would think microsoft.com is presumably just as vunerable to domain theft as joespixoftoilets.com. Also, while if the folks who did this to Panix were Aussies, and if they tried it with IBM or Microsoft, those companies would have lawyers literally pounding on the door of the offending registrar within minutes, who's to say the person can't do it at some ISP in Russia or some other country where the likelyhood of having any legal weight is practically nil?
I worked at a bank in the early 90s that had a trading system based on SQL Server and the client was written in Visual Basic 3. Apart from every other bad design choice in this system (I inherited it when the designers got promoted and started working on another, even bigger system), the all important record counter was an integer, so when trade 32768 was posted, the application crashed, and simply could not be started again, because the first thing it did was try to show the current total (it was written for operators to use, not traders). Worse was that the counter variable wasn't a global, and it was often times a stack variable, and always with a different name (sometimes iCounter, sometimes iCount, sometimes x).
The upshot was that I was able to convince management to totally scrap it and allow me to write a new one. The downside was that the idiot who designed the original system went on to spend 100 million dollars on this new, grandious system that too was eventually scrapped, but he knew long before that his turkey wasn't going to fly, so he quit and became a lead architect at some other company.
*Sigh*...okay, back to coding.
IBM sells a product simply called VM. Actually, I guess it's more of a lease option, as it's only available for mainframes (and I used it on an ES/9000...one of the biggest mainframe (read: MVS/CICS) machines around). It's cool in that you could assign separate processors separate copies of the OS, unlike VMware which has a "host" operating system and then various Guests. There's still a bit of low-level software, but for us it was seemless (which, given how much everything associated with this machine cost, had better have been).
... strange times. Full of promises yet to be fulfilled. But as someone else pointed out, now that OS X is essentialy Unix, there would be precious little reason to go back to the "personality" scheme. I rather think they'd bring out some kick-ass server type box running multiple copies of OS X server, if that is in fact what they're trying to do.
Interestingly, this brought to mind the Pink operating system that IBM and Apple were working on way-back-when(tm). The idea, if I remember correctly, was to have a low level OS kernel that could run multiple personalities...they talked about a MacOS personality (back when System 8 was still being developed), OS/2 and probably some flavor of Unix.
I remember being at what I believe was the last Unix Convention at the Javits Center in NYC around '92 or '93 and they (IBM) had a prototype Power box that purported to be running a super super early pre-alpha version of it. The guy standing by it wouldn't let me touch it, and all he said he could do was run a "DIR" on what was supposed to be the OS/2 personality (no Mac one in sight, for the obvious reason there never was one). He also mentioned that there was a second box, but they couldn't get it to boot.
*Sigh*
I was actually under the impression it was just going to be a dual core PPC, but I RTFA off os OSNews.com a couple of days ago and I don't really remember it.
Just like those M$ bastards to buy a company like this and then say they're going to stop making a Mac and a Linux version.
Oh, wait...nevermind...
I have a BFA in film from a school that has a couple of entries in the Jargon file, but that's about it. I am totally self-taught in terms of programming, and I've now been a Windows/Unix C++ programmer for seven years.
I started by writing Excel 4 macros, moved up to VB, then went to C, later C++. I also learned a lot of SQL, where I can now write sprocs in T-SQL for Sybase/SQL Server, and PL/SQL for Oracle.
In every "move up", it was seeing a need for the next level up when the stuff at the previous level wasn't good enough (needed custom VBXs for VB, etc.) It wasn't lying, we really did need this stuff, and I was willing to step up to the plate to do it. I put in a *lot* of long hours. I read and read and read and played every single day. In all of 1996, I took July 4th, Thanksgiving, and Christmas off...I worked all the others.
I was also upfront and made mistakes (especially when threads got involved), but I worked like hell to fix, and more importantly, *totally* understand the actual issues. I think it also helped that we didn't have things like the STL to use, so I had to roll my own binary trees and linked lists by hand...a great way to learn how to use pointers.
I'm not so sure this method will work for anyone else, as everyone has their own journey to take. But I worked with a lot of others who had various art degrees who were also in IT, and they did it basically the same way as me...lots of grunt work to understand what was going on, and could demonstrate it at both the interview and on the job.
So I would say that it doesn't matter. I also will say that as one who has interviewed a lot of people, I never take into consideration the "Education" section of the resume. You either know the answer to my questions, or can at least try to reason a good guess, or you don't. You can have a PhD in CompSci from Harvard, but if you can't tell me how to update a table in the database, then it's not going to help.
While it's true that Microsoft will always go for lock in, and their own language makes that easier (C#), I hate to admit it, but their 7.1 C++ compiler is quite good, and compiles every wacky template I throw at it. It's supposedly been able to compile Andresceu's Loki library, as well as Boost++, both very taxing on compilers. I think this is more a case of Herb Sutter's championing than anything else. So while you're probably right, at least their C++ compiler does kick ass.
.DBF level) to put into Access 2.0, and then promply announced that FoxPro was dead. The FoxPro community howled long and hard and MS has basically had to keep it alive ever since, always as a separate product, but it's gotten a lot of tweaks and upgrades to make it look like other MS products (read: tabs).
For the best example of outspoken developers keeping a product alive, look no further than FoxPro. MS bought Foxbase to get their Rushmore technology (which I believe was bitmapped indexes brought to the
This is another product you won't hear pass anyone's lips from MS, but it wouldn't surprise me if we see a FoxPro 9 somewhere +/- the time VS2005 is released.
Yeah, my mistake...it was a PS/2 Model 95 with a 60mhz Pentium chip. Heh...that thing ran Netware 3.12 and serviced both IPX and TCP/IP for a couple hundred people at my school as well as being both a mac/pc file server and handling I think at least 3 printers. It probably ran extra hot from being at 100% cpu utilization all the time.
Ah, you're right. My bad.
Of course, the original PPro's (60/66mhz) were hotter than a 40 watt light bulb. I made the mistake of touching just the side of one when working on a PS/2 that had one. That chip was the G5 of its time...how are they *ever* going to make that thing into a portable?
Heh...mentioning MCA reminded me that there was a board available for Model 95s that turned the machine into an extremely small mainframe, but could still run all MVS apps (including CICS and DB/2). Dunno about the IO on the board, but I guess it was IBMs acknowledgement that it could be done (I believe it was a "proof of concept" that never actually made it into the channels).
.com, they didn't even consider it (of course, it's not like anybody can afford an ES/9000...though I suppose the zSeries is pretty good), and as such they never had a solution that gave acceptable performance.
As far as IO, I worked at a bank that had an ES/9000...one of the biggest around. Thing was half the entire floor of the building, in all its water cooled glory. The way it computed your balance was to go through the VSAM files, find you, then start from the very first entry and add/subtract each payment/debit until it reached the end, and you had your balance. Sure it would cache it for awhile (I think a couple of days), but eventually it'd do the whole add/subtract thing all over again.
I could run execute the cobol program from the web, and then get the result back via MQSeries. Never saw it take more than a second...when I was at a dos console, I'd hit enter and it'd bang back the answer without delay. Fastest thing I'd ever seen.
When I went to another job where they were facing a similar problem (having to find/compute records in massive files) I suggested that a mainframe was the way to go. Being a
No, you're right, and I thought about qualifying that in the post. I would not want to try (actually, maybe I would like to try) to process GE's payroll on an emulated IBM 1401 or 360.
I think it was more the fact that it *can* be run; that the ability to run CICS, once only available on a 3090 or whatever, could also run on a PC running a mainframe emulator is what I was thinking.
But you're right, the only thing that performs as good as a mainframe is another mainframe.
I googled RETAIN and from the couple of sites that I read, it jogged the memory and I realize that I think I'd been on the receiving end of it...I worked at a government agency which had an AS/400 back in '91 and one day our local IBM rep shows up with a new 9 track drive and says he's going to install it. I said I don't think anyone had mentioned it, and he replied that the AS/400 had detected a problem and sent a message to IBM. Now I knew why there was that extra phone line attached it it.
I think RETAIN was/is behind those cool radio-based units IBM reps carried...they had a 2 line LCD screen, a chicklet-style qwerty keyboard, but it was like an early Blackberry...the rep could find out where he had to go next, what the problem was, what parts were necessary, etc. I thought it was *very* *very* cool back then (actually, it's still cool), when I was still struggling to get a 9600bps modem working with Procomm.
Because these books talk about a time when punch cards were still all the rage, and because my Ask Slashdot article was rejected, I'll ask here:
Does anyone still use punch cards? I know some states used punch cards for the Nov 2 election, but I'm wondering if there are still decks of cards at companies waiting to be run through and the output printed on green-n-white paper.
It's not a criticism or putdown question, I can believe there are some jobs on some equipment that just can't (or won't) be ported to something newer, and "what worked for us back then works for us now."
Just curious.
I don't think I ever read this book (born in 1970), but flipping through the pages, it makes me realize what computers still mean to my folks; batch cards, mag tapes, green-n-white printouts.
Therein lies the rub; to my folks, any computer that can be fit in a single box and doesn't live in a raised-floor room, is a toy. It's actually very black and white for them..."yes it's all very nice what those toys can do for the movies, but it takes a *computer* to process GE's payroll."
It also reminds me of when a friend of mine brought his dad in to work to show him what he did. His dad was a serious old school programmer for custom chips for Navy jets. He looked it too...checkered shirt, crew cut, pocket protector (first time I'd ever seen one). My friend shows him the *cough* Powerbuilder app we'd be working on, with its buttons and datawindows, etc., and his dad just went *pft* and waved his hand.
The fact that I can run emulators of any of those systems and they run 10x faster has never made a dent in my folks opinion. As far as they can see, and as far as my friend's dad can see, we're just playing with toys.
Anyone else had that happen?
True enough, Renderman is the standard as far as I know. But because Pixar is both producer and consumer of said software (they also make some other tools that I'm forgetting the names of at the moment), they also have access to the developers, so if the Nemo team wanted water to look "just so", they'd merely have to walk over to the appropriate person's office, and maybe if it wasn't in there before, hey, a little coding and there you go. Even a paying customer like Disney wouldn't get that kind of service.
I have an original clamshell iBook with 64meg of ram...far too small in terms of resources, I found, for OS X. Linux (specifically YDL 3) ran acceptably on it and am able to do a lot more with it than what I could do under OS 9 (ironically, a lot of stuff that winds up running on OS X). Gnome doesn't run *great* on it, but it's acceptable. Plus with YDL it was easy to configure the Airport card in it, so I can sit anywhere and use it.
I think you'll find most people run Linux on older macs to revive and get some more use out of older hardware...I wish I hadn't thrown out my 68k-based Mac so I could try NetBSD on it.
I could also see where someone might have a very specific need to run in 64-bit *now*, instead of waiting for Tiger next year; then a G5-specific Linux kernel like YDL 4 would fit the bill.