C and C++ is not the same. And actually as I see it Microsoft has officially "abandoned" their C compiler, they will not add C99 support and not do much development at all for all I know (unless they changed their mind recently).
True, but there's only one MSVC compiler.exe to do both and I expect the backend code generation path is virtually the same for both (if not actually the same). In the context of this thread I meant "native code compiler" as opposed to.NET compiler anyway.
Actually I do. IIRC the big 3.4 change was the new C++ parser not the ABI - the major ABI change was 4.0, but there have been subtle API changes with most 3.x and 4.x releases. 4.2 had several features over 4.1, notably tuning for Core 2 Duo processors and OpenMP, but was unfortunately slower so everyone skipped it. etc.
I still don't understand why you / tjstork think MS neglect their C/C++ compilers though. OK I can't find SPEC numbers to show either way but I can't think of any major features GCC's C++ compiler has over MSVC's. Conversely MSVC already has LTO and IMO an easier/better preprocessed header implementation. Not that it makes any difference to me; I use GCC on Linux and Solaris and MSVC on Windows anyway.
gcc and egcs come to mind about "changing compilers"
Yes, I know about GCC and EGCS. EGCS was a political fork of GCC that got merged back in at (IIRC) GCC 2.95. EGCS was still GCC at heart. So that's hardly a significant change.
IIRC Red Hat pushed EGCS because it had 64-bit SPARC support. I can't remember any other significant reason EGCS was used over GCC.
This is a strange way of putting it. Why not simply say, "IE: 107 seconds, Firefox: 3.9 seconds"??? I guess they thought more digits is sexier.
The sunspider benchmark outputs its results in ms not seconds so I expect they just copy/pasted.
Back when the U.S. and E.U. governments were suing Microsoft, and Microsoft was trying to defend why it was "impossible" to produce Windows 98 without Internet Exploder.
So the other guy said. I don't understand why they'd claim that in the anti-trust suit - it would show competitive advantage and weaken their position. Citation needed.
Remember those anti-trust cases with the Win95/98 and IE bundling? Yeah, around that time.
No, I don't remember performance being mentioned in the anti-trust cases. And why would they? It would highlight an uncompetitive advantage and weaken their position.
It feels like the.NET koolaid is coming even to the IE team. Microsoft's.NET push now borders on maniacal, standardizing on.NET and in places where it should not be standardized. Performance matters, particularly when processors aren't getting any faster, just more parallel. Microsoft's has left C++ to languish, has all but abandoned C, and as such has no real performance tool in their own arsenal.
But IE isn't built on.NET is it? And there are improvements in MSVC in VS2008 for both C and C++ and they've had OpenMP and a much improved STL for two versions now.
For my interest, when have major OSS projects "changed compilers" to respond? I can't think of any examples.
I don't see that the things they mention are fair or informative tests. Yes, there's some browser infrastructure involved but other components are doing most of the work:
papervision3d.org is entirely down to the Flash plugin
a 3D Java render is entirely down to the Java plugin
sunspider - OK, fair enough, we've known about speed problems with string concatenation in IE since sunspider appeared
ACID - yes, this isn't a priority for this release for IE so this isn't news either
Maybe Firefox 3.1 is much faster than IE 8 but this article doesn't tell me anything new.
Yeah I could care less about the tracking system myself, I still use an old MS Intellimouse ball mouse. No batteries to replace, no lag, and good enough tracking.
but you do need to de-gunk the internal rollers every few weeks. That's one thing I don't miss about ball mice.
the only advantage i can see is that it works on more surfaces than laser mice can. but so do conventional optical mice, which can already go up to 1600 dpi.
Actually TFA says that optical was worse than laser on the tricky surfaces:
A very high-end Razer Boomslang CE optical mouse had trouble maintaining a smooth cursor. An Ideazon Reaper Edge laser mouse faired pretty well, but not perfectly. This just goes to show that laser mice and the Explorer are able to handle difficult surfaces more cleanly than optical mice.
Proprietary is not an advantage. Patented is not an advantage.
That's not from the article per-se, it's from MS's press release that the article quotes. The article tells you it's a quote and uses a different font even so it's pretty obvious.
Patented and proprietary are advantages to MS's investors. That's why they put it in the press release.
If you are in US or UK, I want to work for you. Will you apply for my work permit?
We've investigated this in the past for the UK and we think it's virtually impossible - or at least extremely difficult - to get a full work permit for a foreign national for an entry-level developer job. If you've just studied here then you can get a year or two to remain in the UK after your degree - I forget what that's called - but after that your best bet is the Highly-Skilled Migrant Programme which you apply for yourself. You basically need to prove that you have a good degree and evidence of earning power.
There are plenty of examples of game-modder hired by a game dev team, which is kind of analogous I suppose. The only real open-source examples I know are people hired to continue working on that project, e.g. Apple hired the LLVM guy so they can build their next compiler on LLVM, a number of GCC devs have been hired by chip devs or GCC consulting firms or by Linux distros or Apple, etc.
Yep - I know there's patches for it, but apparently MSFT hadn't bothered to push it via Windows Update... I think they're kinda torn between wanting to sell 2k7 licenses and trying to push the format.)
Actually Office 2003 will give you an explanation prompt and the URL to download the format convertors from when you try to open an Office 2007 document.
The bootloader you're talking about is the MBR (the other one is at the start of the active partition). The MBR is 512 bytes, which also includes the partition table.
In short: No, there is not nearly enough space there.
Partitions begin on cylinder boundaries not the next sector. You've got the whole rest of the first cylinder to use too.
on windows swap file always gets used but in terms of the general rule of 2x the ram you have, now days you don't, i have 4gb ram and i been using 2gb swap for years with 0 problems
The only issue with using less swap than RAM on Windows is that you can't do a full memory dump on blue screen because in the first instance it writes the dump to the swap area.
That said, if you didn't know that you probably don't need it. The mini-dump it creates is usually good enough for most crash analysis.
When Eric Clapton plays a mean blues guitar solo, he can do so minus the countless medallions, bling jewelry, fashion street clothing and any need to wave his hands at the camera.
But then there's Hendrix who burns his guitar on stage or one of the countless others who smashes theirs up at the end of the set. The posturing isn't the music.
Blues - yes, you've got a point, but it isn't necessarily everyone's thing. You could say the same about jazz or classical - it's out there, there's an incredible history of it, but they aren't pop culture today.
...to learn to play an instrument and write your own music, just steal someone else's, talk all over it and call it "rap".
Rapping well actually has a lot of skills in common with 'real' singing - it's more technical that you'd think. For example you need at least as much timing, enunciation and breath control as singers do, and if you're rapping tunefully you've got to actually hit the note first time, you can't fudge it in vibrato if you're having an off day. Don't dis the rap stars.
Why not? We're not talking about the receptionist assessing them, we're talking about the receptionist giving them the first written test. It will be marked by or discussed with the technical people later.
There's no reason a receptionist can't handle the whole first section of an interview:
welcome the candidate
get them comfortable in a quiet meeting room (bring them a drink, smalltalk, etc.)
process their travel expenses
give them a quick tour of the office if appropriate (minimum toilets, maybe fire exit)
and so forth. And they can then say "I'm sorry but your interviewer is tied up and will be another 15 minutes. In the meantime he asked if you'd start thinking about this. He'll be in to discuss it with you as soon as he can".
But that can go the other way too. Without reference drivers to build proof-of-concepts how do you know that VIA's hardware actually does everything you need in the first place? If VIA have no reference drivers to give you how did they test it themselves - how far can you trust them? Does the hardware contain bugs that your drivers would expose? You wouldn't discover that until months into development when it'd cost a fortune to change chip. Support lifetime, OK, fair enough - you'd need to get that in the contract. If anyone else has been using the Intel drivers then they'll have gone through the initial debugging and you can have better confidence in them. If VIA want me to take on risk doing their dev for them I'd need a substantial discount.
In the end we're probably arguing over something that will never happen anyway - they'll never just leave it *all* to FOSS.
The end user isn't the person buying a single unit, it's the person buying a few tens of thousands to ship in their systems. These people don't need to rely on volunteers, they hire someone to write the drivers from the documentation.
I'm not sure that helps. Hypothetically:
Intel say "We can supply you 100,000 chips at 50 cents each, including working drivers from day one and software support." VIA say "We can supply you with 100,000 chips at 40 cents each, 10 cents cheaper than Intel. We estimate it will cost you $100,000 and three months to develop your own drivers from our documentation."
Intel are cheaper and - even if they weren't - have no risk of delaying your project. So whose do you buy?
How about Option #3: Don't provide reference code at all, just properly document the hardware and less the FOSS crowd write their own?
In addition to the other replies, there's this reason - you're putting the end-user first impression, which is absolutely critical for PR and repeat business, in someone else's hands. You can't assume volunteers will deliver to your product launch cycles. You can't even assume you'll get competent and motivated volunteers at all - you probably will, but I wouldn't bet the farm on it.
Slashdot Mirror
C and C++ is not the same. And actually as I see it Microsoft has officially "abandoned" their C compiler, they will not add C99 support and not do much development at all for all I know (unless they changed their mind recently).
True, but there's only one MSVC compiler .exe to do both and I expect the backend code generation path is virtually the same for both (if not actually the same). In the context of this thread I meant "native code compiler" as opposed to .NET compiler anyway.
Microsoft say they haven't done full C99 because of lack of user interest, although they note that much of C99 is required for compatibility with the next C++ standard. Mind you if you want C99 with an MSVC backend you can always use Comeau C.
You apparently don't follow the history of GCC:
Actually I do. IIRC the big 3.4 change was the new C++ parser not the ABI - the major ABI change was 4.0, but there have been subtle API changes with most 3.x and 4.x releases. 4.2 had several features over 4.1, notably tuning for Core 2 Duo processors and OpenMP, but was unfortunately slower so everyone skipped it. etc.
I still don't understand why you / tjstork think MS neglect their C/C++ compilers though. OK I can't find SPEC numbers to show either way but I can't think of any major features GCC's C++ compiler has over MSVC's. Conversely MSVC already has LTO and IMO an easier/better preprocessed header implementation. Not that it makes any difference to me; I use GCC on Linux and Solaris and MSVC on Windows anyway.
gcc and egcs come to mind about "changing compilers"
Yes, I know about GCC and EGCS. EGCS was a political fork of GCC that got merged back in at (IIRC) GCC 2.95. EGCS was still GCC at heart. So that's hardly a significant change.
IIRC Red Hat pushed EGCS because it had 64-bit SPARC support. I can't remember any other significant reason EGCS was used over GCC.
This is a strange way of putting it. Why not simply say, "IE: 107 seconds, Firefox: 3.9 seconds"??? I guess they thought more digits is sexier.
The sunspider benchmark outputs its results in ms not seconds so I expect they just copy/pasted.
Back when the U.S. and E.U. governments were suing Microsoft, and Microsoft was trying to defend why it was "impossible" to produce Windows 98 without Internet Exploder.
So the other guy said. I don't understand why they'd claim that in the anti-trust suit - it would show competitive advantage and weaken their position. Citation needed.
Remember those anti-trust cases with the Win95/98 and IE bundling? Yeah, around that time.
No, I don't remember performance being mentioned in the anti-trust cases. And why would they? It would highlight an uncompetitive advantage and weaken their position.
Citation needed.
It feels like the .NET koolaid is coming even to the IE team. Microsoft's .NET push now borders on maniacal, standardizing on .NET and in places where it should not be standardized. Performance matters, particularly when processors aren't getting any faster, just more parallel. Microsoft's has left C++ to languish, has all but abandoned C, and as such has no real performance tool in their own arsenal.
But IE isn't built on .NET is it? And there are improvements in MSVC in VS2008 for both C and C++ and they've had OpenMP and a much improved STL for two versions now.
For my interest, when have major OSS projects "changed compilers" to respond? I can't think of any examples.
I don't see that the things they mention are fair or informative tests. Yes, there's some browser infrastructure involved but other components are doing most of the work:
Maybe Firefox 3.1 is much faster than IE 8 but this article doesn't tell me anything new.
This destroys Microsoft's claim that their intimate knowledge of the OS that runs IE will increase performance.
Oh? When did they ever claim that?
Yeah I could care less about the tracking system myself, I still use an old MS Intellimouse ball mouse. No batteries to replace, no lag, and good enough tracking.
but you do need to de-gunk the internal rollers every few weeks. That's one thing I don't miss about ball mice.
the only advantage i can see is that it works on more surfaces than laser mice can. but so do conventional optical mice, which can already go up to 1600 dpi.
Actually TFA says that optical was worse than laser on the tricky surfaces:
A very high-end Razer Boomslang CE optical mouse had trouble maintaining a smooth cursor. An Ideazon Reaper Edge laser mouse faired pretty well, but not perfectly. This just goes to show that laser mice and the Explorer are able to handle difficult surfaces more cleanly than optical mice.
Proprietary is not an advantage. Patented is not an advantage.
That's not from the article per-se, it's from MS's press release that the article quotes. The article tells you it's a quote and uses a different font even so it's pretty obvious.
Patented and proprietary are advantages to MS's investors. That's why they put it in the press release.
In Windows, you can set the folder to read-only
I don't think read-only folder flag actually does anything.
Try using the security tab to revoke your own permissions to the folder.
If you are in US or UK, I want to work for you. Will you apply for my work permit?
We've investigated this in the past for the UK and we think it's virtually impossible - or at least extremely difficult - to get a full work permit for a foreign national for an entry-level developer job. If you've just studied here then you can get a year or two to remain in the UK after your degree - I forget what that's called - but after that your best bet is the Highly-Skilled Migrant Programme which you apply for yourself. You basically need to prove that you have a good degree and evidence of earning power.
There are plenty of examples of game-modder hired by a game dev team, which is kind of analogous I suppose. The only real open-source examples I know are people hired to continue working on that project, e.g. Apple hired the LLVM guy so they can build their next compiler on LLVM, a number of GCC devs have been hired by chip devs or GCC consulting firms or by Linux distros or Apple, etc.
ARM has, to my knowledge, only been used in embedded systems.
Actually ARM's first ever use was in a desktop. But they don't make them any more.
Yep - I know there's patches for it, but apparently MSFT hadn't bothered to push it via Windows Update... I think they're kinda torn between wanting to sell 2k7 licenses and trying to push the format.)
Actually Office 2003 will give you an explanation prompt and the URL to download the format convertors from when you try to open an Office 2007 document.
Partitions begin on cylinder boundaries not the next sector. You've got the whole rest of the first cylinder to use too.
on windows swap file always gets used but in terms of the general rule of 2x the ram you have, now days you don't, i have 4gb ram and i been using 2gb swap for years with 0 problems
The only issue with using less swap than RAM on Windows is that you can't do a full memory dump on blue screen because in the first instance it writes the dump to the swap area.
That said, if you didn't know that you probably don't need it. The mini-dump it creates is usually good enough for most crash analysis.
When Eric Clapton plays a mean blues guitar solo, he can do so minus the countless medallions, bling jewelry, fashion street clothing and any need to wave his hands at the camera.
But then there's Hendrix who burns his guitar on stage or one of the countless others who smashes theirs up at the end of the set. The posturing isn't the music.
Blues - yes, you've got a point, but it isn't necessarily everyone's thing. You could say the same about jazz or classical - it's out there, there's an incredible history of it, but they aren't pop culture today.
...to learn to play an instrument and write your own music, just steal someone else's, talk all over it and call it "rap".
Rapping well actually has a lot of skills in common with 'real' singing - it's more technical that you'd think. For example you need at least as much timing, enunciation and breath control as singers do, and if you're rapping tunefully you've got to actually hit the note first time, you can't fudge it in vibrato if you're having an off day. Don't dis the rap stars.
... are not given a test by a receptionist.
Why not? We're not talking about the receptionist assessing them, we're talking about the receptionist giving them the first written test. It will be marked by or discussed with the technical people later.
There's no reason a receptionist can't handle the whole first section of an interview:
get them comfortable in a quiet meeting room (bring them a drink, smalltalk, etc.)
and so forth. And they can then say "I'm sorry but your interviewer is tied up and will be another 15 minutes. In the meantime he asked if you'd start thinking about this. He'll be in to discuss it with you as soon as he can".
But that can go the other way too. Without reference drivers to build proof-of-concepts how do you know that VIA's hardware actually does everything you need in the first place? If VIA have no reference drivers to give you how did they test it themselves - how far can you trust them? Does the hardware contain bugs that your drivers would expose? You wouldn't discover that until months into development when it'd cost a fortune to change chip. Support lifetime, OK, fair enough - you'd need to get that in the contract. If anyone else has been using the Intel drivers then they'll have gone through the initial debugging and you can have better confidence in them. If VIA want me to take on risk doing their dev for them I'd need a substantial discount.
In the end we're probably arguing over something that will never happen anyway - they'll never just leave it *all* to FOSS.
The end user isn't the person buying a single unit, it's the person buying a few tens of thousands to ship in their systems. These people don't need to rely on volunteers, they hire someone to write the drivers from the documentation.
I'm not sure that helps. Hypothetically:
Intel say "We can supply you 100,000 chips at 50 cents each, including working drivers from day one and software support."
VIA say "We can supply you with 100,000 chips at 40 cents each, 10 cents cheaper than Intel. We estimate it will cost you $100,000 and three months to develop your own drivers from our documentation."
Intel are cheaper and - even if they weren't - have no risk of delaying your project. So whose do you buy?
This is not a new concept (how many MMO's have given exclusive items to people who pre-order?)
Yeah, but it's normally something cosmetic like a pet or a different look armor. I can't think of any examples that affect gameplay.
How about Option #3: Don't provide reference code at all, just properly document the hardware and less the FOSS crowd write their own?
In addition to the other replies, there's this reason - you're putting the end-user first impression, which is absolutely critical for PR and repeat business, in someone else's hands. You can't assume volunteers will deliver to your product launch cycles. You can't even assume you'll get competent and motivated volunteers at all - you probably will, but I wouldn't bet the farm on it.