I bet food in a normal supermarket is handled and mishandled by more bacteria-ridden people than the warehouse of the delivery service.
All you have to do is judge the quality of the food you receive from the online place. If it sucks, don't buy there. For many, it's convenient. Several people I know use a similar service here and are very happy with it.
90/10 is not a law or empirical result, just a rule of thumb, and probably rarely true within a reasonable enough margin of error to plug it into another equation. So yes, 10 is just pulled out of someone's ass.
Obviously some codes scale far beyond an order of magnitude or there wouldn't be supercomputers with hundreds or thousands of processors.
Amdahl's law was used as an argument against massively parallel processing in the 1960s because it shows dramatically diminishing returns. Presumably IBM wanted people to buy their single processor systems... (See also Gustafson's law from the 1980s) It wouldn't apply to true parallel algorithms, for example. As other posters stated, the constant factor in Amdahl's law can change that curve significantly.
But something like this could replace X in years to come
I think it would take something amazing to replace the X Window System. There have been a few attempts to replace it. picoGUI appears to be a inspired by QNX's Photon microGUI (hence the blatantly similar name).
I think you hit on the real problem though, standardization. There is no single central Unix vendor than can dictate UI design standards or enforce a single widget toolkit. picoGUI has a snowball's chance in hell of becoming the defacto widget/windowing system. Hence it will just serve to fracture the UI experience even more.
My complaint about the UI on Linux/Unix has always been that it's so frickin inconsistent. Too often apps do things differently. Not better, just differently. Everyone can make their own super neato widget library and it's not that hard to make your own X replacement.
That's the problem. Don't try to invent your own window system, menu ordering or UI interactions, just base them on Windows as much as possible and forget about "the best" way to do things. With KDE (don't know about Gnome, never liked it) that is happening, slowly. Maybe picoGUI is taking the right approach by limiting the ability to make apps with fucked up GUIs just because you can and "Windows sucks", but most Unix diehards will loathe the loss of flexibility.
This suggests that writing extensions is so hard that only a select few can do it. Not encouraging
The RandR paper mentions the graveyard of failed X extensions. X extensions should be a last resort IMO, so it doesn't bother me if there isn't a new one every week.
Yes, I should have specified that I was talking about integer performance. I'm not sure what percentage of typical user tasks benefit from SIMD.
Altivec is supposed to be more flexible. However, Altivec is not perfect either - it doesn't work on double precision floating point values like SSE2 does as I understand it.
Since the other poster answered your question, I just wanted to make the comment that the GNU Hurd blurs this user/kernel space distinction by minimizing what is in the kernel. A user can run and test their own kernel components independently from other users.
No apologies necessary. I don't mind if things are closed. After all my, programs are all closed source. I was just trying to get across that neither PDF or XDoc is a standard like an ISO standard, that is vendor-neutral, and I wasn't that clear. Of course I do like using published de facto file formats and protocols at a minimum and open standards because it seems more flexible in the ling run.
I'm not creating my own definitions here. There are just multiple definitions for "open". Are you familiar with The Open Group? I don't think you're qualified to discuss "the industry" as though you were a seasoned expert on standards issues if you are ignorant of that fact.
Bear in mind that my original use of "open" was immediately preceded by me stating that Adobe controlled it and certainly a difference in definitions was no justification to call me an idiot, just like your haughtiness is unjustified.
I personally use the term "open standard" for standards not controlled by vendors, and "published standards" otherwise.
You sound like you're parroting some XML marketing blurb. Well, I have been working with XML for years and I actually know what I'm talking about from experience and having read many of the W3C XML specs.
I totally disagree with it being a matter of "knowing what the tags are". It's not that simple. What about CDATA and other encodings? What about the fragility of the format? Others have already posted markup from Microsoft that has hex strings. What the hell does a hex string mean? You need to be able to decode it. PostScript is usually plain text. Have you ever looked at a PostScript file with embedded font information? You'd have to reverse engineer that data unless you knew the binary font format.
No, PDF is just not an editing format. I'm not alking about read/write. Sure you _can_ edit it, but it's not a working document format. Working document formats carry semantic information like "this is a paragraph".
I can't see how Xdocs gives MS any more leverage than they already have with.doc,.xls and so on.
It gives them a lot of leverage. PDF users are a significant market segment, especially in high end publishing. If MS's format becomes popular then they gain leverage over Adobe and other PDF supporters, and consequently the publishing industry. It appears that XDocs is primarily for forms, however, which is just one use of PDFs (and I've never been thrilled about them as I've said elsewhere).
Addendum. If you mean free as in "does it require a license fee to create PDFs like Unisys and their evil GIFs"? No it is free beer as far as licensing.
Portable Document Format. As in documents you can edit.
Not true! PDF is designed as a display/rendering format, not an editable document format. Generally SGML or XML content is rendered into PDF and then PDF is rendered into output.
Slashdot Mirror
All you have to do is judge the quality of the food you receive from the online place. If it sucks, don't buy there. For many, it's convenient. Several people I know use a similar service here and are very happy with it.
-Kevin
Obviously some codes scale far beyond an order of magnitude or there wouldn't be supercomputers with hundreds or thousands of processors.
Amdahl's law was used as an argument against massively parallel processing in the 1960s because it shows dramatically diminishing returns. Presumably IBM wanted people to buy their single processor systems... (See also Gustafson's law from the 1980s) It wouldn't apply to true parallel algorithms, for example. As other posters stated, the constant factor in Amdahl's law can change that curve significantly.
-Kevin
4 megadominos
-Kevin
I think it would take something amazing to replace the X Window System. There have been a few attempts to replace it. picoGUI appears to be a inspired by QNX's Photon microGUI (hence the blatantly similar name).
I think you hit on the real problem though, standardization. There is no single central Unix vendor than can dictate UI design standards or enforce a single widget toolkit. picoGUI has a snowball's chance in hell of becoming the defacto widget/windowing system. Hence it will just serve to fracture the UI experience even more.
My complaint about the UI on Linux/Unix has always been that it's so frickin inconsistent. Too often apps do things differently. Not better, just differently. Everyone can make their own super neato widget library and it's not that hard to make your own X replacement.
That's the problem. Don't try to invent your own window system, menu ordering or UI interactions, just base them on Windows as much as possible and forget about "the best" way to do things. With KDE (don't know about Gnome, never liked it) that is happening, slowly. Maybe picoGUI is taking the right approach by limiting the ability to make apps with fucked up GUIs just because you can and "Windows sucks", but most Unix diehards will loathe the loss of flexibility.
-Kevin
The RandR paper mentions the graveyard of failed X extensions. X extensions should be a last resort IMO, so it doesn't bother me if there isn't a new one every week.
-Kevin
-Kevin
The clock speed rating ratio between P4s and G4s is very close to the performance difference.
-Kevin
-Kevin
-Kevin
Multics had multithreading in the early 1970s.
Windows was still launched from DOS in 1992.
Please go back to your "innovating" with Windows.
-Kevin
-Kevin
Bear in mind that my original use of "open" was immediately preceded by me stating that Adobe controlled it and certainly a difference in definitions was no justification to call me an idiot, just like your haughtiness is unjustified.
I personally use the term "open standard" for standards not controlled by vendors, and "published standards" otherwise.
-Kevin
I totally disagree with it being a matter of "knowing what the tags are". It's not that simple. What about CDATA and other encodings? What about the fragility of the format? Others have already posted markup from Microsoft that has hex strings. What the hell does a hex string mean? You need to be able to decode it. PostScript is usually plain text. Have you ever looked at a PostScript file with embedded font information? You'd have to reverse engineer that data unless you knew the binary font format.
-Kevin
-Kevin
-Kevin
No, PDF is just not an editing format. I'm not alking about read/write. Sure you _can_ edit it, but it's not a working document format. Working document formats carry semantic information like "this is a paragraph".
Right, I meant open as in controlled by a standards organization, not open as in free of licensing (which it is).
I usually use gv in Linux for viewing pdfs which AFAIK doesn't allow cut & paste.
-Kevin
Ack you're right. I've gone insane. Sorry about that.
-Kevin
It gives them a lot of leverage. PDF users are a significant market segment, especially in high end publishing. If MS's format becomes popular then they gain leverage over Adobe and other PDF supporters, and consequently the publishing industry. It appears that XDocs is primarily for forms, however, which is just one use of PDFs (and I've never been thrilled about them as I've said elsewhere).
-Kevin
Addendum. If you mean free as in "does it require a license fee to create PDFs like Unisys and their evil GIFs"? No it is free beer as far as licensing.
-Kevin
Not true! PDF is designed as a display/rendering format, not an editable document format. Generally SGML or XML content is rendered into PDF and then PDF is rendered into output.
-Kevin
-Kevin
PDF is widely used in the printing industry. It allows metadata and can be used in a workflow.
-Kevin