Naah, I was joking. I just like the idea of Gonzalez giving the Bush war cabinet (i.e. Bush and Barney) insane legal advice that just happened to be what Bush wanted to hear as a way to be the AG.
I think 'unwritten' is self deprecation or cynicism. It's true in the sense that there isn't one document with a small set of authors that describes the British system. That doesn't mean that you can piece together a constitution from the sources you descibe though. Mind you that constitution would be very complex and not at all logical.
Though as a Tory and programmer I think it's like a very old piece of code which has been patched for a long time, hard to understand but for good reasons. Certainly the English system has a lot of staying power. It's been tested by much worse things than the current Islamist threat and it has survived. Other simpler systems might not be as lucky.
Oh wait, I forgot - they're not being held by the police, and they're not actually in America. My bad. They're also not people, at least according to legal advice given by Alberto Gonzalez verbally to the Bush and his terrier Barney in a cold war bunker under his ranch at 4:32am on 9/11/2001. Gonzalez didn't just promised to forget the conversation, he actually did. By 4:35am, he had really had forgotten.
But this dubious but very convenient piece of legal mindfuck, mostly inspired by Babylonian Numerology and obesseive reading of lurid pulp biographies of famous tyrants whilst high on cocaine rather than more traditional American legal sources, lead to Gonzalez being nominated for Attorney General 3 years later.
There his record was at best mixed, but that is another story.
If you're respinning your ASICs you're not doing it right. You're also wasting $500k to $millions per spin, which would be far better spent on verification. I guess if you're picking a random roomful of people, and think it's a simple task to integrate and verify a cpu core or other IP, then trial and error is pretty much your only choice. Us professionals prefer to do it right the first time.
Call bullshit all you want Mr. Armchair ASIC Designer. When you've taped out a nanometer-level ASIC you'll have an opinion on the subject worth listening to. Oh fuck off. I'm not an armchair designer, that description fits you far better with statements like 'us professionals prefer to get it right first time'. If you're so shit hot why do you waste time telling strangers about it on the internets? Don't you have better things to do, like save companies millions by getting it right first time. Fucking poser. I have worked with companies that have done it, not with random people but the best ones they could hire, which usually means rather mediocre people. And it's not that hard, just expensive.
Re:Anything else out there?
on
The State of X.Org
·
· Score: 2, Insightful
It's slow because you have to aptget unfree-leachware-closed-source to get an accelerated driver which will probably fuck the machine up.
As opposed to a closed source driver from ATI that fucks your machine up on Windows?
ATI aren't that bad. My old machine had an ATI card and it was rock solid. I prefer NVidia's drivers though, the ATI.Net bloated control panel annoys me.
NVidia Windows drivers always seem to mature faster than ATI's for new cards or new Windows versions, usually they are stable in a couple of revisions. ATI sometimes go through bad patches for driver stability.
Or even better one of those noname brand cards that have better support on Linux than the crappy drivers on the CD that come with the card?
I don't buy cards like that.
2D acceleration is enabled for all major brands of video cards (ATI, Nvidia, Via, and Intel [and probably Matrox too, but I don't have a card to test it on]) in the default free Xorg drivers. Sometimes the X configurator in the distro decides to use a generic framebuffer for no good reason. It's easy to fix.
Really? Last time I tried it didn't support my card, the latest NVidia. Given that it takes people time to reverse engineer hardware in the free driver I suspect that the latest card will always be like that.
It was actually pretty hard to get X out of 320*200 256 colors, non accelerated. Eventually I nuked the system and put XP back on it.
And lets face it, if I have the latest and greatest graphics card, I really want accelerated 3D, not just 2D. Not that there are any games for X I'd actually play of course. But it seems like one of the stengths of Windows is that most games target DirectX. The drivers are stable and you can get fast hardware. The fact that DirectX is supported for most games mean the graphics cards are optimized to implement it efficiently. Of course, the NVidia Windows driver has accelerated OpenGL too.
All of which means that a Windows PC is still a tempting target for videogames.
I can Ctrl+C and Ctrl+V to cut'n'paste between applications.
The Windows clipboard beats any clipboard KDE or Gnome has. This is primarily an application level failure, since X doesn't even know what applications put on the clipboard (and none of the applications know what another application has put on, either). Just throwing a mime-type on the clipboard blob would probably do wonders for KDE/Gnome clipboard functionality. The content negotiation X supports is just not specific enough (and it's old formats).
See it's the "X doesn't know what the applications put on the clipboard" thing that makes me hate X. Clipboards are not a recent invention. The problem with X Windows is that its creators seems to have this hippy view of the world where the Windowing system shouldn't impose choices on application writers but rather leave them free to make their own choices. Fuck that, I want a machine with a fascist conformity where every application conforms to the style "guidelines" or it is sent to a digital death camp by brownshirts like me.
And the thing is most users are fascists like me. 90% buy Windows machines where applications have a high degree of conformity. Another 3% buy Macs which are even more fascist.
Now it used to be that X Windows had network transparency and Windows didn't. But now there's Remote Desktop and VNC to handle that. And it was a far better idea to optimize for the common case where the application is running locally than across the network.
Remote Desktop is not truly network transparency; it's just remote desktop. Same with VNC. X lets you run applications anywhere and view them anywhere.
Yeah, but I don't need that network transparency, I only need remote desktop. And having it there means you can't optimize for the common case where video memory is local. On Windows there are lots of neat ways for applications to get
No. It's much harder. Rockets are large, visible with the naked eye, and adhere to macro-level physics and mechanics. Nanometer-scale semiconductor design involves a lot more work to avoid many, many more possible modes of failure. Meh, what a load of bullshit.
I've had several clients that have done this stuff. They prototyped on a shitload of FPGAs and then moved to an ASIC built by a couple of fabs. And these were not simple designs. You need a roomful of people and a lot of cash - since it takes a few respins of the ASIC to get something usable and a respin is not cheap - but it can be done by an averagely disorganised medium sized company with decidedly average people. Most stuff you license e.g CPU cores, the rest you design in house.
Or you could pay someone in Taiwan to do the whole lot, since they have a lot of experience in this stuff.
And if you want to make money out of games consoles, you have to do this.
To fix the problem, Microsoft went back to an unnamed ASIC vendor based in the United States and redesigned the chip, Lewis added. (Based on a previous report, the ASIC vendor is most likely the former ATI Technologies, now part of AMD.)
The funny thing about this is: ATI is not an ASIC vendor! ATI does chip design and, since they're fabless (or were until AMD bought them,) they get them made at TSMC or sometimes Chartered, or, most often, use NEC Electronics Amertica as their ASIC vendor. ATI partnered with Microsoft to make an ASIC (first at TSMC, then later with NEC) but, in generall, you don't go to ATI to get an ASIC made. You go to NEC, or IBM, or Toshiba, or LSI Logic, etc. Those are ASIC vendors. ATI is a fabless design house specializing in graphics. Big difference. EETimes seems to think they are.
I know companies that are a lot smaller than Microsoft who've done ASICs and it has worked.
Without an ASIC vendor? As in, taped out GDS2 directly to a fab like TSMC? What process node? If you say 90nm or lower (which is the kind of ASIC we're talking about here) I'd have to call bullshit and ask for the names of these companies. Can't comment on that, NDAs and client confidentiality.
Re:Anything else out there?
on
The State of X.Org
·
· Score: 2, Insightful
So what is the important stuff that the Windows display manager does, and X doesn't? Umm not suck? I hate having to use X. It's slow because you have to aptget unfree-leachware-closed-source to get an accelerated driver which will probably fuck the machine up. Cut and paste never works. Keyboard accelerators don't work. All the applications look different. The fonts are too small and are ugly. It's fail, complete and utter.
Whereas on Windows I have a nice accelerated display by default. I can Ctrl+C and Ctrl+V to cut'n'paste between applications. Decent windows applications have accelerators so I don't need to move my hands off the keyboard. And because Windows is a look and feel rather than just a toolkit I can see how to use a Windows app pretty much instantly. And Windows has had True Type since 3.1 so the fonts look right to me.
Now it used to be that X Windows had network transparency and Windows didn't. But now there's Remote Desktop and VNC to handle that. And it was a far better idea to optimize for the common case where the application is running locally than across the network. It meant that 16 Windows apps could run quite well on Windows 3.1 on a 286 at 12Mhz.
These days a 32 or 64 bit Windows app in C or C++ is seriously simple to write and it very efficient since you build the UI out of built in Windows classes. You subclass 'em and override the 1% of behaviour you want to change. And since you're only overriding 1%, everyone knows how to use it. You can make small exe files too, e.g. uTorrent.
There is a good alternative to X Windows, it's called Win32. From what I've read uTorrent on Wine is still pretty lightweight.
Basically he's trying to create business for ASIC design houses by telling people that putting a bunch of licensed IP onto a chip is rocket science and they shouldn't try to do it in house.
Is it really? I honestly don't know. I suspect it depends a lot on the quality of the in house people and the quality of the ASIC design house.
It is true. You should not unnecessarily muck with VHDL/Verilog and 3rd party cores even if you work with an FPGA. This will not kill you, but it will make you poorer. HDLs are notoriously kludgy, and it takes a lot of effort to do it right. Proprietary cores rarely work as documented, and you have no visibility into them. When multiple cores are used, it's one large fingerpointing game between vendors. And you need to have good, experienced HDL coders. And you need to have all the tools, they cost big bucks.
But that's with mere FPGAs, where you can update your design whenever you wish. However here they are talking about ASICs - where all the wiring is done with masks when the IC is made. You'd have to be certifiably mad to even think about a casual design like this. ASIC designs are done by very competent teams, using "10% coding / 90% verification" time allocation, because you can't afford/any/ mistakes. And even then you make mistakes; but experienced teams with good tools make those mistakes smaller, and they call them "errata" - something that is not right but can be worked around. When you make the F0 0F bug, though, you trash the whole run.
I know companies that are a lot smaller than Microsoft who've done ASICs and it has worked.
So Microsoft risked a lot when it went for an in-house design. I am not surprised that they failed. They should have counted all the successful 3D video companies on the market and asked themselves why there are so few, and why top gaming cards cost so much.
But if you're a cash rich company then the bias will be to try to do as much as possible in house, because that gives you more freedom to value engineer later.
I am not MS, but I don't really see much business value in rolling your own video controller. More likely the NIH syndrome kicked in, or some people were overly concerned about their job security.
Yeah, but they didn't. They licensed the IP from ATI. Whether it was VHDL or a hard core I don't know. But the whole Xenos chip is an ATI design. Microsoft had the chip layout done in house and talked to TSMC directly.
http://www.eetimes.com/showArticle.jhtml;jsessionid=IB2CS0JQTKT1SQSNDLOSKH0CJUNN2JVN?articleID=199902345 Lewis pointed to the example of Microsoft which licensed for its Xbox 360 a graphics design from the former ATI Technologies and had it made at TSMC. The company considered a similar move for its Xbox 360 processor designed by IBM, but at the last minute decided to have IBM assume overall responsibility for making, packaging and testing the chip rather than buying raw wafers from Chartered Semiconductor.
When that didn't work they got ATI to do the layout.
http://www.eetimes.com/news/semi/showArticle.jhtml?articleID=208403010 To fix the problem, Microsoft went back to an unnamed ASIC vendor based in the United States and redesigned the chip, Lewis added. (Based on a previous report, the ASIC vendor is most likely the former ATI Technologies, now part of AMD.)
But now they presumably know the secret. Which will come in handy whem they want to start reducing the chipcount so they can cut the losses they make selling the things and/or cut the price.
So ASIC design is risky, but it means that XBox360's are no longer sold at a loss
I use this article pretty regularly in presentations to point out flaws in the Open Source model, and as a prime example of idiocy itself.
End user's I deal with care allot about "workflow". Perhaps software on the home PC is about making people happy - at least for the small [yes, it is small] portion of the population that spends time every day on social networking sites and the like. But most of the software used every day is for the purpose of doing work.
I guess he would think a mail client that integrates with MySpace is a killer app. I dunno, I think he was criticising soulless software like Lotus Notes. And he was right, Netscape was fucked around this point. He's funny too, unlike the drones that mumble buzzwords like "enterprise" and "groupware". They were boring bureaucrats, only interested in making money from other boring bureaucrats. And in the end they didn't even manage that. So what use were they?
You have to like someone who says this
"Groupware" is all about things like "workflow", which means, "the chairman of the committee has emailed me this checklist, and I'm done with item 3, so I want to check off item 3, so this document must be sent back to my supervisor to approve the fact that item 3 is changing from 'unchecked' to 'checked', and once he does that, it can be directed back to committee for review."
Nobody cares about that shit. Nobody you'd want to talk to, anyway.
Which really is the point. Even if you can make money out of "that shit", do you really want to? When you were at University, did you really think you'd be a replaceable cog in a big machine, with an average salary (until your job is outsourced or right sized away), producing a inferior clone of Lotus Notes so that people could file their TPS reports online? Fuck that. I wanted to be a Pirate Of Silicon Valley or a video game programmer. At any rate someone who would either make it big through Godlike business and coding skill or crash and burn spectacularly.
Ah well, I suppose I should let you go back to filling in your WebTPS report. If you ever decide to go postal, please kill the people in my rival department, not me ktxbye.
You know, the more I read Jamie Zawinski, the more I wonder what the fuck I'm doing as an engineer in a large company. Consider.
http://www.jwz.org/doc/groupware.html Now the problem here is that the product's direction changed utterly. Our focus in the client group had always been to build products and features that people wanted to use. That we wanted to use. That our moms wanted to use.
"Groupware" is all about things like "workflow", which means, "the chairman of the committee has emailed me this checklist, and I'm done with item 3, so I want to check off item 3, so this document must be sent back to my supervisor to approve the fact that item 3 is changing from 'unchecked' to 'checked', and once he does that, it can be directed back to committee for review."
Nobody cares about that shit. Nobody you'd want to talk to, anyway.
Users GOOD
If you want to do something that's going to change the world, build software that people want to use instead of software that managers want to buy.
When words like "groupware" and "enterprise" start getting tossed around, you're doing the latter. You start adding features to satisfy line-items on some checklist that was constructed by interminable committee meetings among bureaucrats, and you're coding toward an externally-dictated product specification that maybe some company will want to buy a hundred "seats" of, but that nobody will ever love. With that kind of motivation, nobody will ever find it sexy. It won't make anyone happy.
Ok, I said it was a funny story, but obviously that's not the funny part, unless sad is funny.
I think he wrote another article on the utter idiocy of rewriting Netscape so the code became nice and easy to read too. In both cases he's basically sick of humouring bright people who have completely lost touch with reality because they are stuck in their own little world of refactoring or business alliances or open source. Anything that convinces bright people that they don't need to solve hard problems, just apply some "magic pixie dust" that will make those hard problems all disappear.
And now he's running a bar. I wonder how long before I am.
Denmark are just part of the general howl of protest from people who've looked at the heap of excrement that is DIS 29500 and found it wanting Yeah, the poor standard of DIS 29500 is a popular topic of conversation where I am. Just now I popped out for a burger and the guy that gave it to me was complaining about blatant violation of ISO voting procedures.
The Microsoft/OOXML spam on Slashdot has gotten as bad as the Ron Paul spam on Digg. OK we get you are fanatical about a subject (ie standards) but please enough is enough. In the scheme of things this MS/OOXML subject is not that important despite what you zealots think. Hey I'm writing software that reads OOXML and what happens to my kids if the standard is nuked and so no one wants to buy it.
That's right, government cheese. In a van. Down by the river.
Also, since the fabric ovviously stretches and such when opening the doors (looking at the video), I'd wonder how long it would keep its original shape, before it would stretch and start flapping & making noise when you're driving down the freeway. The skin on my BMW hangs loose, like Wizard's sleeve. Maybe I should trade it in for a newer model.
You are still misunderstanding. No one said OSX is dead. MacOS died at version 9, and OSX replaced it. OSX is a *NIX based operating system. Thus the statement that only UNIX and Windows remain. Mac GUI applications still use the Mac specific APIs rather the X Windows ones though. Describing OS X as Unix is disingenous because while OS X can run Unix applications, both GUI and command line, Unix can't run Mac GUI applications.
All sufficiently complex systems must also be incomplete. Lawyers have another name for this "the law is an ass".
In a democracy the incompleteness of the law isn't a problem because the legislature can keep adding laws to patch up the most obvious holes. Though I suppose you could argue that what you're talking about is patching up too.
But I think the difference at least with English law to software is that it's not designed to be objectively readable. The law says something and the CPS interpret it before deciding to try someone. But they jury might decide to interpret it differently. The meaning of the language can even change retroactively depending on case law. All this stuff is far too subtle for current machines to check. But a democratic society can do it. If the law is deficient so that the outcome of court cases strikes people as unjust the legislature can patch it.
Income of $524 million Loss of $423 million Equals a spread of $947 million, almost 1 billion.
Add the $1 billion recall, still looks like Vista and Office are paying for the XBOX 360. Yup, just goes to show you that engineering elegance is not the same as profitability.
I've seen groups crushed and the politicians move out before the axe falls. The problem is, they led the charge straight to hell and only got off the bus at the toll charge. Ouch. This is a unique combination of a mixed metaphor (a charge to hell on a bus where the leader can get off?) and using charge to mean too different things in one sentence. Say what you like about MBAs, but I bet they wouldn't write like this.
;-) (Joking since I regularly write stuff that is flat out incomprehensible gibberish)
I don't think you're getting it. Cutting costs is one thing. Cutting corners is another. Cutting costs is fine, but cutting corners implies the product is worse off because of it. Few engineers would say "It'd be cheaper to roll our own graphics chip," because they realize the immense technical challenges involved. They didn't "roll their own graphics chip" from what I can tell. They licensed the IP (the VHDL code or a synthesized core) from someone else. The plan from the start with the XBox360 was that they would do this and try to integrate it all eventually onto one chip. That's the reason they moved from x86 to PPC, because neither Intel or AMD would license their IP and let Microsoft make their own chips. Actually this is the difference between Risc and x86 these days - x86 vendors don't license their IP but Risc vendors do. Since consoles are sold at a loss initially and subsidized by games it's really important to reduce the build costs by doing this. Back in the XBox days most people thought that Microsoft lost out because they couldn't integrate the design into once chip in the way that Sony did with their console. And that was because they didn't own the IP for the processor.
The mistake seemed to be to let Microsoft's in house group do this rather than outsourcing.
Basically he's trying to create business for ASIC design houses by telling people that putting a bunch of licensed IP onto a chip is rocket science and they shouldn't try to do it in house.
Is it really? I honestly don't know. I suspect it depends a lot on the quality of the in house people and the quality of the ASIC design house.
And it depends on what you're trying to do. In the embedded area lots of companies much smaller than Microsoft put an processor and a bunch of their own peripherals onto a chip and it works. I guess that console or PC graphics cores use a lot more power than that. But I don't know if "an ASIC design house" would have done a better job than Microsoft's ASIC group.
Or more to the point, maybe a $1B recall is the price you pay for learning about this stuff. Microsoft can afford it obviously and it will influence how the successor to the XBox360 is done. Whether they hire more engineers and do it in house or outsource it is a business decision it seems. I guess the in house people and the design house will both try to argue for the best option from their point of view and some manager will decide.
But if you're a cash rich company then the bias will be to try to do as much as possible in house, because that gives you more freedom to value engineer later.
I dunno, the problem I have with MBA types as managers is that it's easier to learn the business stuff yourself than the technology.
And for balance the problem I have with engineers as managers is that it's possible to learn the people skills stuff but you have to understand why it's important and want to do learn it. It's all too easy to stay in the comfort zone where you basically sit in a dark corner somewhere and write code if that's what you enjoy rather than forcing yourself to talk to people.
If you read the other comments I made in this thread what I'm doing is not the same as the 'creation science' people. I only believe things if they can survive the sort of experimental battering Relativity did and if they can additionally explain things that current physics can't.
Slashdot Mirror
Naah, I was joking. I just like the idea of Gonzalez giving the Bush war cabinet (i.e. Bush and Barney) insane legal advice that just happened to be what Bush wanted to hear as a way to be the AG.
I think 'unwritten' is self deprecation or cynicism. It's true in the sense that there isn't one document with a small set of authors that describes the British system. That doesn't mean that you can piece together a constitution from the sources you descibe though. Mind you that constitution would be very complex and not at all logical.
Though as a Tory and programmer I think it's like a very old piece of code which has been patched for a long time, hard to understand but for good reasons. Certainly the English system has a lot of staying power. It's been tested by much worse things than the current Islamist threat and it has survived. Other simpler systems might not be as lucky.
... where it's currently 6+ years and counting.Oh wait, I forgot - they're not being held by the police, and they're not actually in America. My bad. They're also not people, at least according to legal advice given by Alberto Gonzalez verbally to the Bush and his terrier Barney in a cold war bunker under his ranch at 4:32am on 9/11/2001. Gonzalez didn't just promised to forget the conversation, he actually did. By 4:35am, he had really had forgotten.
But this dubious but very convenient piece of legal mindfuck, mostly inspired by Babylonian Numerology and obesseive reading of lurid pulp biographies of famous tyrants whilst high on cocaine rather than more traditional American legal sources, lead to Gonzalez being nominated for Attorney General 3 years later.
There his record was at best mixed, but that is another story.
Call bullshit all you want Mr. Armchair ASIC Designer. When you've taped out a nanometer-level ASIC you'll have an opinion on the subject worth listening to. Oh fuck off. I'm not an armchair designer, that description fits you far better with statements like 'us professionals prefer to get it right first time'. If you're so shit hot why do you waste time telling strangers about it on the internets? Don't you have better things to do, like save companies millions by getting it right first time. Fucking poser. I have worked with companies that have done it, not with random people but the best ones they could hire, which usually means rather mediocre people. And it's not that hard, just expensive.
It's slow because you have to aptget unfree-leachware-closed-source to get an accelerated driver which will probably fuck the machine up.
As opposed to a closed source driver from ATI that fucks your machine up on Windows?
ATI aren't that bad. My old machine had an ATI card and it was rock solid. I prefer NVidia's drivers though, the ATI .Net bloated control panel annoys me.
NVidia Windows drivers always seem to mature faster than ATI's for new cards or new Windows versions, usually they are stable in a couple of revisions. ATI sometimes go through bad patches for driver stability.
Or even better one of those noname brand cards that have better support on Linux than the crappy drivers on the CD that come with the card?
I don't buy cards like that.
2D acceleration is enabled for all major brands of video cards (ATI, Nvidia, Via, and Intel [and probably Matrox too, but I don't have a card to test it on]) in the default free Xorg drivers. Sometimes the X configurator in the distro decides to use a generic framebuffer for no good reason. It's easy to fix.
Really? Last time I tried it didn't support my card, the latest NVidia. Given that it takes people time to reverse engineer hardware in the free driver I suspect that the latest card will always be like that.
It was actually pretty hard to get X out of 320*200 256 colors, non accelerated. Eventually I nuked the system and put XP back on it.
And lets face it, if I have the latest and greatest graphics card, I really want accelerated 3D, not just 2D. Not that there are any games for X I'd actually play of course. But it seems like one of the stengths of Windows is that most games target DirectX. The drivers are stable and you can get fast hardware. The fact that DirectX is supported for most games mean the graphics cards are optimized to implement it efficiently. Of course, the NVidia Windows driver has accelerated OpenGL too.
All of which means that a Windows PC is still a tempting target for videogames.
I can Ctrl+C and Ctrl+V to cut'n'paste between applications.
The Windows clipboard beats any clipboard KDE or Gnome has. This is primarily an application level failure, since X doesn't even know what applications put on the clipboard (and none of the applications know what another application has put on, either). Just throwing a mime-type on the clipboard blob would probably do wonders for KDE/Gnome clipboard functionality. The content negotiation X supports is just not specific enough (and it's old formats).
See it's the "X doesn't know what the applications put on the clipboard" thing that makes me hate X. Clipboards are not a recent invention. The problem with X Windows is that its creators seems to have this hippy view of the world where the Windowing system shouldn't impose choices on application writers but rather leave them free to make their own choices. Fuck that, I want a machine with a fascist conformity where every application conforms to the style "guidelines" or it is sent to a digital death camp by brownshirts like me.
And the thing is most users are fascists like me. 90% buy Windows machines where applications have a high degree of conformity. Another 3% buy Macs which are even more fascist.
Now it used to be that X Windows had network transparency and Windows didn't. But now there's Remote Desktop and VNC to handle that. And it was a far better idea to optimize for the common case where the application is running locally than across the network.
Remote Desktop is not truly network transparency; it's just remote desktop. Same with VNC. X lets you run applications anywhere and view them anywhere.
Yeah, but I don't need that network transparency, I only need remote desktop. And having it there means you can't optimize for the common case where video memory is local. On Windows there are lots of neat ways for applications to get
I've had several clients that have done this stuff. They prototyped on a shitload of FPGAs and then moved to an ASIC built by a couple of fabs. And these were not simple designs. You need a roomful of people and a lot of cash - since it takes a few respins of the ASIC to get something usable and a respin is not cheap - but it can be done by an averagely disorganised medium sized company with decidedly average people. Most stuff you license e.g CPU cores, the rest you design in house.
Or you could pay someone in Taiwan to do the whole lot, since they have a lot of experience in this stuff.
And if you want to make money out of games consoles, you have to do this.
The funny thing about this is: ATI is not an ASIC vendor! ATI does chip design and, since they're fabless (or were until AMD bought them,) they get them made at TSMC or sometimes Chartered, or, most often, use NEC Electronics Amertica as their ASIC vendor. ATI partnered with Microsoft to make an ASIC (first at TSMC, then later with NEC) but, in generall, you don't go to ATI to get an ASIC made. You go to NEC, or IBM, or Toshiba, or LSI Logic, etc. Those are ASIC vendors. ATI is a fabless design house specializing in graphics. Big difference. EETimes seems to think they are.
Without an ASIC vendor? As in, taped out GDS2 directly to a fab like TSMC? What process node? If you say 90nm or lower (which is the kind of ASIC we're talking about here) I'd have to call bullshit and ask for the names of these companies. Can't comment on that, NDAs and client confidentiality.
Whereas on Windows I have a nice accelerated display by default. I can Ctrl+C and Ctrl+V to cut'n'paste between applications. Decent windows applications have accelerators so I don't need to move my hands off the keyboard. And because Windows is a look and feel rather than just a toolkit I can see how to use a Windows app pretty much instantly. And Windows has had True Type since 3.1 so the fonts look right to me.
Now it used to be that X Windows had network transparency and Windows didn't. But now there's Remote Desktop and VNC to handle that. And it was a far better idea to optimize for the common case where the application is running locally than across the network. It meant that 16 Windows apps could run quite well on Windows 3.1 on a 286 at 12Mhz.
These days a 32 or 64 bit Windows app in C or C++ is seriously simple to write and it very efficient since you build the UI out of built in Windows classes. You subclass 'em and override the 1% of behaviour you want to change. And since you're only overriding 1%, everyone knows how to use it. You can make small exe files too, e.g. uTorrent.
There is a good alternative to X Windows, it's called Win32. From what I've read uTorrent on Wine is still pretty lightweight.
Basically he's trying to create business for ASIC design houses by telling people that putting a bunch of licensed IP onto a chip is rocket science and they shouldn't try to do it in house.
Is it really? I honestly don't know. I suspect it depends a lot on the quality of the in house people and the quality of the ASIC design house.
It is true. You should not unnecessarily muck with VHDL/Verilog and 3rd party cores even if you work with an FPGA. This will not kill you, but it will make you poorer. HDLs are notoriously kludgy, and it takes a lot of effort to do it right. Proprietary cores rarely work as documented, and you have no visibility into them. When multiple cores are used, it's one large fingerpointing game between vendors. And you need to have good, experienced HDL coders. And you need to have all the tools, they cost big bucks.
But that's with mere FPGAs, where you can update your design whenever you wish. However here they are talking about ASICs - where all the wiring is done with masks when the IC is made. You'd have to be certifiably mad to even think about a casual design like this. ASIC designs are done by very competent teams, using "10% coding / 90% verification" time allocation, because you can't afford /any/ mistakes. And even then you make mistakes; but experienced teams with good tools make those mistakes smaller, and they call them "errata" - something that is not right but can be worked around. When you make the F0 0F bug, though, you trash the whole run.
I know companies that are a lot smaller than Microsoft who've done ASICs and it has worked.
So Microsoft risked a lot when it went for an in-house design. I am not surprised that they failed. They should have counted all the successful 3D video companies on the market and asked themselves why there are so few, and why top gaming cards cost so much.
But if you're a cash rich company then the bias will be to try to do as much as possible in house, because that gives you more freedom to value engineer later.
I am not MS, but I don't really see much business value in rolling your own video controller. More likely the NIH syndrome kicked in, or some people were overly concerned about their job security.
Yeah, but they didn't. They licensed the IP from ATI. Whether it was VHDL or a hard core I don't know. But the whole Xenos chip is an ATI design. Microsoft had the chip layout done in house and talked to TSMC directly.
http://www.eetimes.com/showArticle.jhtml;jsessionid=IB2CS0JQTKT1SQSNDLOSKH0CJUNN2JVN?articleID=199902345
Lewis pointed to the example of Microsoft which licensed for its Xbox 360 a graphics design from the former ATI Technologies and had it made at TSMC. The company considered a similar move for its Xbox 360 processor designed by IBM, but at the last minute decided to have IBM assume overall responsibility for making, packaging and testing the chip rather than buying raw wafers from Chartered Semiconductor.
When that didn't work they got ATI to do the layout.
http://www.eetimes.com/news/semi/showArticle.jhtml?articleID=208403010
To fix the problem, Microsoft went back to an unnamed ASIC vendor based in the United States and redesigned the chip, Lewis added. (Based on a previous report, the ASIC vendor is most likely the former ATI Technologies, now part of AMD.)
But now they presumably know the secret. Which will come in handy whem they want to start reducing the chipcount so they can cut the losses they make selling the things and/or cut the price.
So ASIC design is risky, but it means that XBox360's are no longer sold at a loss
End user's I deal with care allot about "workflow". Perhaps software on the home PC is about making people happy - at least for the small [yes, it is small] portion of the population that spends time every day on social networking sites and the like. But most of the software used every day is for the purpose of doing work.
I guess he would think a mail client that integrates with MySpace is a killer app. I dunno, I think he was criticising soulless software like Lotus Notes. And he was right, Netscape was fucked around this point. He's funny too, unlike the drones that mumble buzzwords like "enterprise" and "groupware". They were boring bureaucrats, only interested in making money from other boring bureaucrats. And in the end they didn't even manage that. So what use were they?
You have to like someone who says this
"Groupware" is all about things like "workflow", which means, "the chairman of the committee has emailed me this checklist, and I'm done with item 3, so I want to check off item 3, so this document must be sent back to my supervisor to approve the fact that item 3 is changing from 'unchecked' to 'checked', and once he does that, it can be directed back to committee for review."
Nobody cares about that shit. Nobody you'd want to talk to, anyway.
Which really is the point. Even if you can make money out of "that shit", do you really want to? When you were at University, did you really think you'd be a replaceable cog in a big machine, with an average salary (until your job is outsourced or right sized away), producing a inferior clone of Lotus Notes so that people could file their TPS reports online? Fuck that. I wanted to be a Pirate Of Silicon Valley or a video game programmer. At any rate someone who would either make it big through Godlike business and coding skill or crash and burn spectacularly.
Ah well, I suppose I should let you go back to filling in your WebTPS report. If you ever decide to go postal, please kill the people in my rival department, not me ktxbye.
No, and I don't write office software. I was joking and wanted an excuse to use the wonderful phrase 'government cheese' because it makes me laugh.
You know, the more I read Jamie Zawinski, the more I wonder what the fuck I'm doing as an engineer in a large company. Consider.
http://www.jwz.org/doc/groupware.html
Now the problem here is that the product's direction changed utterly. Our focus in the client group had always been to build products and features that people wanted to use. That we wanted to use. That our moms wanted to use.
"Groupware" is all about things like "workflow", which means, "the chairman of the committee has emailed me this checklist, and I'm done with item 3, so I want to check off item 3, so this document must be sent back to my supervisor to approve the fact that item 3 is changing from 'unchecked' to 'checked', and once he does that, it can be directed back to committee for review."
Nobody cares about that shit. Nobody you'd want to talk to, anyway.
Users GOOD
If you want to do something that's going to change the world, build software that people want to use instead of software that managers want to buy.
When words like "groupware" and "enterprise" start getting tossed around, you're doing the latter. You start adding features to satisfy line-items on some checklist that was constructed by interminable committee meetings among bureaucrats, and you're coding toward an externally-dictated product specification that maybe some company will want to buy a hundred "seats" of, but that nobody will ever love. With that kind of motivation, nobody will ever find it sexy. It won't make anyone happy.
Ok, I said it was a funny story, but obviously that's not the funny part, unless sad is funny.
I think he wrote another article on the utter idiocy of rewriting Netscape so the code became nice and easy to read too. In both cases he's basically sick of humouring bright people who have completely lost touch with reality because they are stuck in their own little world of refactoring or business alliances or open source. Anything that convinces bright people that they don't need to solve hard problems, just apply some "magic pixie dust" that will make those hard problems all disappear.
And now he's running a bar. I wonder how long before I am.
That's right, government cheese. In a van. Down by the river.
All sufficiently complex systems must also be incomplete. Lawyers have another name for this "the law is an ass".
In a democracy the incompleteness of the law isn't a problem because the legislature can keep adding laws to patch up the most obvious holes. Though I suppose you could argue that what you're talking about is patching up too.
But I think the difference at least with English law to software is that it's not designed to be objectively readable. The law says something and the CPS interpret it before deciding to try someone. But they jury might decide to interpret it differently. The meaning of the language can even change retroactively depending on case law. All this stuff is far too subtle for current machines to check. But a democratic society can do it. If the law is deficient so that the outcome of court cases strikes people as unjust the legislature can patch it.
Congress shall make no law exceeding in length this Constitution.
(Let's make them earn their pay by holding a separate vote on every pork-laden amendment.) ptriot act section 56889
fbi n cia can spy on ppl. dnt need to ask cngress only prez. prez can amend ptriot act with exec order. prez is root. sudo ptriot act.
Loss of $423 million
Equals a spread of $947 million, almost 1 billion.
Add the $1 billion recall, still looks like Vista and Office are paying for the XBOX 360. Yup, just goes to show you that engineering elegance is not the same as profitability.
The mistake seemed to be to let Microsoft's in house group do this rather than outsourcing.
But you've got to remember this is an article in EEtimes from an analyst with an agenda
http://www.eetimes.com/news/latest/showArticle.jhtml;jsessionid=51TYZYXYRWUZUQSNDLSCKHA?articleID=208403010
"System OEMs have no business designing ASICs any longer," said Lewis. The reality is that system companies are finding it hard to do enough ASIC designs to keep in-house design teams employed.
Basically he's trying to create business for ASIC design houses by telling people that putting a bunch of licensed IP onto a chip is rocket science and they shouldn't try to do it in house.
Is it really? I honestly don't know. I suspect it depends a lot on the quality of the in house people and the quality of the ASIC design house.
And it depends on what you're trying to do. In the embedded area lots of companies much smaller than Microsoft put an processor and a bunch of their own peripherals onto a chip and it works. I guess that console or PC graphics cores use a lot more power than that. But I don't know if "an ASIC design house" would have done a better job than Microsoft's ASIC group.
Or more to the point, maybe a $1B recall is the price you pay for learning about this stuff. Microsoft can afford it obviously and it will influence how the successor to the XBox360 is done. Whether they hire more engineers and do it in house or outsource it is a business decision it seems. I guess the in house people and the design house will both try to argue for the best option from their point of view and some manager will decide.
But if you're a cash rich company then the bias will be to try to do as much as possible in house, because that gives you more freedom to value engineer later.
I dunno, the problem I have with MBA types as managers is that it's easier to learn the business stuff yourself than the technology.
And for balance the problem I have with engineers as managers is that it's possible to learn the people skills stuff but you have to understand why it's important and want to do learn it. It's all too easy to stay in the comfort zone where you basically sit in a dark corner somewhere and write code if that's what you enjoy rather than forcing yourself to talk to people.
Did you try to say that to bullies in school?
Not at all, whether I believe a theory or not is not based on that.
If you read the other comments I made in this thread what I'm doing is not the same as the 'creation science' people. I only believe things if they can survive the sort of experimental battering Relativity did and if they can additionally explain things that current physics can't.