I was talking of the first microcomputer, but he wasn't producted "en masse", just done by a guy in his garage.
Presumably that antedated the Micral, as the page I found on the Computer Museum site seems to imply that the Micral wasn't just a single machine made in a garage (although maybe the first Micral was built in said garage).
(To bring in a reference to another Frenchman's work, was it an airtight garage?:-))
For the first microprocessor i always thought this was Intel who did it. Is that right?
I have the impression that the Intel 4004 was, in fact, the first single-chip microprocessor.
I seem to remember seeing something that indicated that the first commercial microprocessor-based computer was French;
Found it. This page from the Computer Museum Web site says that "The Micral was the earliest commercial, non-kit personal computer based on a micro-processor, the Intel 8008."
I also seem to remember the name "Philippe Kahn" (yes, that Philippe Kahn) associated with it.
It also says "Thi Truong developed the computer and Philippe Kahn the software."
Microprocessor, or computer built around a microprocessor? I seem to remember seeing something that indicated that the first commercial microprocessor-based computer was French; I also seem to remember the name "Philippe Kahn" (yes, that Philippe Kahn) associated with it.
Re:Well, a guy I know says...
on
Which BSD?
·
· Score: 2
I asked him to elaborate on this compatibility thing, and he said "Well... I think... OpenBSD can run C++ programs." Instantly I lost all respect for him. I inquired further, and he said "yes, it can run Microsoft C++ programs, and the other BSD's can't."
Perhaps he had it backwards; OpenBSD is the BSD that concentrates on security, and, as for the Microsoft programs, the "About Wine" page on the Wine Web site mentions Linux, FreeBSD, and Solaris (presumably meaning Solaris x86) as platforms on which Wine can run Windows x86 binaries, but doesn't mention openBSD.
...which can make life difficult for awhile until you figure out why a perfectly good command doesn't work anymore...
You don't need to have used Linux to have that experience - you can experience it with nothing but a mix of different commercial UNIXes, for example.:-) (Or you can experience it moving to a Linux system, just as you can experience it moving from a Linux system, or you can experience it moving to a BSD system from a commercial UNIX system, or from a commercial UNIX system to a BSD system.)
"UNIX systems all behave the same, except where they don't."
(I suspect you can also experience it moving from Windows 9x to Windows NT, or vice versa....)
There is no sparc port of FreeBSD. FreeBSD runs only on i386 and alpha platforms.
There was, at one point, a SPARC port project, but I think it may have died out (the link to it on the Projects page on the FreeBSD Web site is broken). (I have the vague impression that Sun was encouraging it for use on, perhaps, some UltraSPARC-based boards they sold, but may have lost interest.)
I think there may also be an IA-64 port in progress.
Re:BSD's deserve a look...
on
Which BSD?
·
· Score: 2
Even M$ uses the BSD tcp/ip stack.
Indeed? I've heard that claimed, but not seen any hard evidence one way or the other.
MS does, as I remember, use, for example, the BSD FTP client (I think I ran "strings" on "ftp.exe" and saw a Berkeley copyright notice in it), but I've not seen anything to indicate that their TCP/IP stack came from BSD.
Re:My impressions of BSD's
on
Which BSD?
·
· Score: 2
For example FreeBSD has ported KDE as of 1.1.1 but NetBSD is up to 1.1.2.
FreeBSD is up to 1.1.2 - that's what I'm running at home. (The NetBSD 1.1.2 ports were mentioned in a news item on the KDE Web site; I didn't see the FreeBSD ones mentioned, so you may not have known about 1.1.2's availability on FreeBSD.)
It's probably just robo-generated text
on
Which BSD?
·
· Score: 2
Who the fuck is this AC?
Probably somebody who picked the "Complain about a company/organization" option for Scott Pakin's automatic complaint-letter generator, supplying "NetBSD" as the name of the company/organization, specified that it should generate 5 paragraphs, hit the "Complain" button, and pasted the results into a Slashdot comment.
Fortunately for the world, I never did get that far
Damn. Any interest in trying again, this time with a Qt 2.0 or GTK+ 1.2 theme?
that's only a short step from talking paperclips...
Anybody know if Microsoft has an SDK for the Office "Assistants"? Having Tux or Beastie suggest you use StarOffice or KOffice or... instead, or having Mr. Hanky the Christmas Poo or Bart Simpson or some other alternative, could be amusing....
My question (which was rhetorical; I already knew the answer to said question was "no") was about KDE 1.x, as I was replying to somebody who said
been able to theme kde for quite a while now..
...
under 1.0 it didn't work well, but 1.1 added a great theme manager, and 1.1.1 built apon that even more.
so the correct answer to my question is not "yes, everything", given that, as you note
because KDE 1.x is based on QT 1.4 it cannot change as many things
The poster to whom I was replying was presumably replying to
When KDE 2.0 comes out, we will be able to really start themeing our look and feel, something the GTK has been much better at in the past.
in another post, a comment that was referring to the ability to theme widgets (as is clear from the stuff following said comment).
The whole point of my comment was that the ability to theme the UI to that extent is new in KDE 2.0; it's not something that people have "been able to [do] for quite a while now", unless they've been running pre-release KDE 2.
I can look like a mac, beos, or whatever i want.. even changing icons, title bars, everything really.
Everything? Including, say, scrollbars? The theming added in Qt 2.0 is the ability to make the widgets look different (beyond the Windows vs. Motif stuff Qt 1.x can do), not just the stuff drawn by programs such as the window manager or file manager.
People talk as if an action taken by a single Christian is representitive of all Christians.
This might be less likely to happen if certain groups didn't appear to use the term "Christian" to mean "conforming to our entire belief system", e.g. this "Christian, Family Oriented ISP", who emphasises that they "are dedicated to protecting our families from pornographic and foul language websites on the Internet" (admittedly, I'm a nonbeliever, but I didn't have the impression that Jesus spent much energy worrying about pornography and foul language).
Perhaps if that particular type of Christian didn't try to divide the world into "Christians" who think as they do and, presumably, "non-Christians" who don't, it might be a bit easier to see Christians as the diverse sort they are, rather than identifying Christianity with fundamentalism. (I also think people who aren't that particular type of Christian should make an effort not to view all of Christianity as being no more than fundamentalism - and should do the same for other religions, e.g. Islam.)
but *if* Intel manages to scale IA-64 way beyond Sparc, I wouldn't think that Sun would have any problem switching over.
At least not any technical problem.:-) There might be a corporate pride problem (look how long it took Scott McNealy to publicly use that "M" word:-)).
The address-space limitations may require a 64-bit processor, but the file size limitations don't, as long as you're not mapping the entire file into your address space - plenty of OSes running on 32-bit processors allow files bigger than 4GB.
Not to mention native support for double precision floating point math in both hardware AND software.
If by "double precision" you mean "64-bit floating-point numbers", with the exponent+mantissa+sign taking 64 bits total, what current 32-bit processors don't support 64 or more bits of floating-point number in hardware (and/or don't run software that supports them as well)?
Integer arithmetic instructions that handle at most 32 bits at a time, and 32-bit pointers, don't limit file sizes to 32 bits (one can synthesize 64-bit or bigger arithmetic operations from the integer arithmetic instructions - and most C compilers, these days, will do it for you, and compilers for other languages may do the same); they only limit how big a file you can map into your address space all at once.
They also don't limit the width of floating-point numbers.
The key limitation of 32-bit processors is the address space limitation you noted; that one is a bit more of a pain to work around (e.g., by mapping stuff into and out of your address space manually).
Remember, the Playstation2 is what, 128bit?
64-bit, or 64-bit with some 128-bit instructions. The Emotion Engine is, according to a Microprocessor Report article, a processor that implements the MIPS III instruction set plus some MIPS IV stuff, with some additional vector and MPEG-2 decoding support added. The operand bus for the vector stuff is 128 bits, but that's not enough to make it a full-blown 128-bit processor (any more than the SIMD stuff in MMX/SSE, or AltiVec, or..., make processors that implement those wider everywhere).
Solaris/UltraSparc is 64 bit; IA64 supports IA32 instructions.
But IA-32 instructions, as the "32" suggests, don't support a 64-bit flat address space, so, to run Solaris 7 on IA-64, Sun has to treat Tterbium or Otassium or whatever the heck it's called as an IA-64 processor; the fact that it runs IA-32 instructions doesn't help (unless you boot Solaris x86 on it, but all that would demonstrate is "yup, Adium's IA-32 mode works").
More like one year, given that this press release has Sun announcing Solaris 7 on October 27, 1998, and given that Solaris 7 was the first fully 64-bit OS from Sun (Solaris 2.6 supports 64-bit file offsets, but that's it).
The title of the original article was a bit misleading, as "64-bit Solaris" has been running for about a year on Sun's SPARC V9 systems; however, I suspect at least some nerds consider it news that it also now boots on Raseodymium or Echnetium or Odium or whatever the hell that name was that Intel presumably spent lots of money to get somebody to come up with....
I suspect not; I don't remember "OS/32" being associated with it, and its file system was UNIXish (and its API UNIX-compatible).
Perhaps you meant Concurrent OS/32? OS/32 was originally a (non-UNIX-compatible) OS from Interdata; Interdata got bought by Perkin-Elmer, was spun off from Perkin-Elmer as Concurrent Computer, and merged with Masscomp (here's a page on their history). It looks as if they're still supporting OS/32, and the machines on which it ran, to some degree, according to the "Latest Facts" page on the Series 3200 on Concurrent's Web site.
(Some of my earliest systems programming work was at a summer job at Interdata, back in 1972; they were located near where my family lived. This was long before OS/32 existed - at the time I was there, some people there were working on an new OS that would add support for disk files that had names.)
A couple months ago a friend mailed me a question... programs were starting to segfault with alarming frequency, and he noticed some obscure message in his kern.log (forget what it was) and he had no idea what was wrong... so I go into/usr/src/linux, do a recursive grep (grep is the best tool ever!) on the error message, read the appropriate code surrounding the message to figure out what's going on, and mail the guy back with some suggestions on how to fix the problem. That is why open source rules, and that is why I love Linux.
Perhaps - although even better might have been a knowledge base (think HOWTO, if the term "knowledge base" sets off anti-Microsoft antibodies:-)) so that one wouldn't have to search the kernel source to find out what that message was saying, whether what it's saying could explain the problems one is seeing, and how to fix the problem.
It's nice that you can open the hood/bonnet and check the engine - sometimes the answer might not be in a knowledge base, and sometimes the answer in the knowledge base might not be the right answer for the particular problem you're seeing, and it's also nice that you can take a look and see what software is doing, see how it's doing it, and see why it's doing it, and it's nice that you can make it do something different - but it's not necessarily nice if you have to do that.
that the question you are arguing so vehemently over is, in a strictly philosophical sense, completely trivial and of no importance whatsoever;)
I'm not sure I'd go quite that far. Yes, it's all speculative, but one's position on what would've happened had Linux not been open-source may influence one's position on whether some project should be, or should become, open-source, and discussion might (or might not, of course...) help one eliminate speculations weakly supported or not supported by what data we have on what did happen with various software projects, and thus might (or might not...) help one make a better decision on said (hypothetical) project.
However, it may be wise to avoid coming up with Final Definite Conclusions about the effect of making projects open-source until there's a large set of projects that did, and didn't, go open-source, and enough time since they began to decide whether they succeeded or failed (bearing in mind, of course, Keynes' comment that, in the long run, we are all dead - or the statement it's claimed Zhou Enlai made when asked about the historical importance of the French Revolution, "It's too early to tell" - how long do you have to wait before you can decide what effect something had, and, by that time, is the answer meaningful or relevant?).
(And, of course, the projects may differ in ways having nothing to do with the openness of the source and that cause them to succeed or fail independently of whether they're open-source or not.)
(What I'd like to know is...where can I get the One True Unix all these variants are varying from?)
Here. (Go for "UNIX V7"; the APIs most modern UNIX systems share with older UNIXes largely look like those of V7, with additions, rather than V6 and the V6-ish PWB/UNIX 1.0; you might want to get PWB to get some of the stuff in the System N releases as well.)
I suppose System V Release 2 or so could also be thought of as the UNIX most commercial UNIXes these days derive from, either directly or indirectly, although SVR2's APIs, as noted, largely resemble a superset of V7, with some PWBisms, and some other additions, e.g. shared memory, and some changes, e.g. a different TTY driver "ioctl" interface.
SCO's only charging USD 100 for the license, it appears - and they also may "waive the license fee for a limited number of deserving applicants", where they suggest that "deserving" may include " student, unemployed, disabled, financially challenged, etc."
The sources are, of course, to PDP-11 UNIXes, so to actually run them you'd need a PDP-11, or a PDP-11 simulator.
But to make broad inroads into the commercial (let alone the consumer) marketplace the command line has to go away.
The OS (or family of OSes) that has the biggest market share in the commercial and consumer marketplace has a command line. It's not as powerful a command line as that of UNIX, perhaps, and it may not be as necessary to drop down to the command line on those OSes as it is on at least some UNIX-flavored desktop OSes, but it does exist.
Give the Pointy-Haired Boss the perception that doing something odd on the command line will recompile the kernel (which, to him, is indistinguishable from issuing a launch order to the Air Force) and he won't touch that computer.
Give the PHB the perception that the command line doesn't exist, and they won't have the perception that there's anything odd to do on the command line.
Till Linux
can't be recompiled by the end user, it won't be competition for Windows.
I see no evidence to support that assertion. I suspect it'd be sufficient, at least at the "recompiling the kernel" level (and, mutatis mutandis, the "recompiling any other part of the system" level), to make it unnecessary to recompile the kernel for most purposes (e.g., making heavy use of loadable kernel modules, etc.).
(There will be applications ("embedded" systems, say, and perhaps at least some servers) where the ability to recompile the kernel may be a win, so, of course, the ability should't disappear.)
Perhaps there will be "end-user desktop" distributions where the source doesn't come on the distribution CD (the GPL doesn't require that what you distribute come with the source, you merely have to make the source available "for a charge no more than your cost of physically performing source distribution", "on a medium customarily used for software interchange"; that medium doesn't have to be the medium that contains the binary).
I'm not sure that's required to calm the fears of PHBs, however. I think the main reason for that might be to take one item off the list of options in the install program, to simplify the installation process, or, if the OS is pre-loaded on machines, it would be done to leave more disk space for pictures of the grandchildren or downloaded music or whatever.
Slashdot Mirror
No.
Presumably that antedated the Micral, as the page I found on the Computer Museum site seems to imply that the Micral wasn't just a single machine made in a garage (although maybe the first Micral was built in said garage).
(To bring in a reference to another Frenchman's work, was it an airtight garage? :-))
I have the impression that the Intel 4004 was, in fact, the first single-chip microprocessor.
Found it. This page from the Computer Museum Web site says that "The Micral was the earliest commercial, non-kit personal computer based on a micro-processor, the Intel 8008."
It also says "Thi Truong developed the computer and Philippe Kahn the software."
Microprocessor, or computer built around a microprocessor? I seem to remember seeing something that indicated that the first commercial microprocessor-based computer was French; I also seem to remember the name "Philippe Kahn" (yes, that Philippe Kahn) associated with it.
Perhaps he had it backwards; OpenBSD is the BSD that concentrates on security, and, as for the Microsoft programs, the "About Wine" page on the Wine Web site mentions Linux, FreeBSD, and Solaris (presumably meaning Solaris x86) as platforms on which Wine can run Windows x86 binaries, but doesn't mention openBSD.
That may, of course, just be because they didn't mention it, not because Wine can't run on OpenBSD; the "emulators" section of the OpenBSD ports status page mentions Wine, so there may be a working port.
You don't need to have used Linux to have that experience - you can experience it with nothing but a mix of different commercial UNIXes, for example. :-) (Or you can experience it moving to a Linux system, just as you can experience it moving from a Linux system, or you can experience it moving to a BSD system from a commercial UNIX system, or from a commercial UNIX system to a BSD system.)
"UNIX systems all behave the same, except where they don't."
(I suspect you can also experience it moving from Windows 9x to Windows NT, or vice versa....)
There was, at one point, a SPARC port project, but I think it may have died out (the link to it on the Projects page on the FreeBSD Web site is broken). (I have the vague impression that Sun was encouraging it for use on, perhaps, some UltraSPARC-based boards they sold, but may have lost interest.)
I think there may also be an IA-64 port in progress.
Indeed? I've heard that claimed, but not seen any hard evidence one way or the other.
MS does, as I remember, use, for example, the BSD FTP client (I think I ran "strings" on "ftp.exe" and saw a Berkeley copyright notice in it), but I've not seen anything to indicate that their TCP/IP stack came from BSD.
FreeBSD is up to 1.1.2 - that's what I'm running at home. (The NetBSD 1.1.2 ports were mentioned in a news item on the KDE Web site; I didn't see the FreeBSD ones mentioned, so you may not have known about 1.1.2's availability on FreeBSD.)
Probably somebody who picked the "Complain about a company/organization" option for Scott Pakin's automatic complaint-letter generator, supplying "NetBSD" as the name of the company/organization, specified that it should generate 5 paragraphs, hit the "Complain" button, and pasted the results into a Slashdot comment.
Damn. Any interest in trying again, this time with a Qt 2.0 or GTK+ 1.2 theme?
Anybody know if Microsoft has an SDK for the Office "Assistants"? Having Tux or Beastie suggest you use StarOffice or KOffice or... instead, or having Mr. Hanky the Christmas Poo or Bart Simpson or some other alternative, could be amusing....
My question (which was rhetorical; I already knew the answer to said question was "no") was about KDE 1.x, as I was replying to somebody who said
so the correct answer to my question is not "yes, everything", given that, as you note
The poster to whom I was replying was presumably replying to
in another post, a comment that was referring to the ability to theme widgets (as is clear from the stuff following said comment).
The whole point of my comment was that the ability to theme the UI to that extent is new in KDE 2.0; it's not something that people have "been able to [do] for quite a while now", unless they've been running pre-release KDE 2.
...even if it runs on other UNIX-compatible OSes?
What's an application developed by folks running Linux, folks running FreeBSD, folks running Solaris, folks running NetBSD, folks running IRIX, ...?
Are you certain that every single line of code in GNOME was developed solely on Linux?
Everything? Including, say, scrollbars? The theming added in Qt 2.0 is the ability to make the widgets look different (beyond the Windows vs. Motif stuff Qt 1.x can do), not just the stuff drawn by programs such as the window manager or file manager.
This might be less likely to happen if certain groups didn't appear to use the term "Christian" to mean "conforming to our entire belief system", e.g. this "Christian, Family Oriented ISP", who emphasises that they "are dedicated to protecting our families from pornographic and foul language websites on the Internet" (admittedly, I'm a nonbeliever, but I didn't have the impression that Jesus spent much energy worrying about pornography and foul language).
Perhaps if that particular type of Christian didn't try to divide the world into "Christians" who think as they do and, presumably, "non-Christians" who don't, it might be a bit easier to see Christians as the diverse sort they are, rather than identifying Christianity with fundamentalism. (I also think people who aren't that particular type of Christian should make an effort not to view all of Christianity as being no more than fundamentalism - and should do the same for other religions, e.g. Islam.)
At least not any technical problem. :-) There might be a corporate pride problem (look how long it took Scott McNealy to publicly use that "M" word :-)).
The address-space limitations may require a 64-bit processor, but the file size limitations don't, as long as you're not mapping the entire file into your address space - plenty of OSes running on 32-bit processors allow files bigger than 4GB.
If by "double precision" you mean "64-bit floating-point numbers", with the exponent+mantissa+sign taking 64 bits total, what current 32-bit processors don't support 64 or more bits of floating-point number in hardware (and/or don't run software that supports them as well)?
Integer arithmetic instructions that handle at most 32 bits at a time, and 32-bit pointers, don't limit file sizes to 32 bits (one can synthesize 64-bit or bigger arithmetic operations from the integer arithmetic instructions - and most C compilers, these days, will do it for you, and compilers for other languages may do the same); they only limit how big a file you can map into your address space all at once.
They also don't limit the width of floating-point numbers.
The key limitation of 32-bit processors is the address space limitation you noted; that one is a bit more of a pain to work around (e.g., by mapping stuff into and out of your address space manually).
64-bit, or 64-bit with some 128-bit instructions. The Emotion Engine is, according to a Microprocessor Report article, a processor that implements the MIPS III instruction set plus some MIPS IV stuff, with some additional vector and MPEG-2 decoding support added. The operand bus for the vector stuff is 128 bits, but that's not enough to make it a full-blown 128-bit processor (any more than the SIMD stuff in MMX/SSE, or AltiVec, or..., make processors that implement those wider everywhere).
But IA-32 instructions, as the "32" suggests, don't support a 64-bit flat address space, so, to run Solaris 7 on IA-64, Sun has to treat Tterbium or Otassium or whatever the heck it's called as an IA-64 processor; the fact that it runs IA-32 instructions doesn't help (unless you boot Solaris x86 on it, but all that would demonstrate is "yup, Adium's IA-32 mode works").
This presumes that there will be "Sun IA-64 stuff", rather than just Sun SPARC hardware. That isn't necessarily going to be the case.
More like one year, given that this press release has Sun announcing Solaris 7 on October 27, 1998, and given that Solaris 7 was the first fully 64-bit OS from Sun (Solaris 2.6 supports 64-bit file offsets, but that's it).
The title of the original article was a bit misleading, as "64-bit Solaris" has been running for about a year on Sun's SPARC V9 systems; however, I suspect at least some nerds consider it news that it also now boots on Raseodymium or Echnetium or Odium or whatever the hell that name was that Intel presumably spent lots of money to get somebody to come up with....
I suspect not; I don't remember "OS/32" being associated with it, and its file system was UNIXish (and its API UNIX-compatible).
Perhaps you meant Concurrent OS/32? OS/32 was originally a (non-UNIX-compatible) OS from Interdata; Interdata got bought by Perkin-Elmer, was spun off from Perkin-Elmer as Concurrent Computer, and merged with Masscomp (here's a page on their history). It looks as if they're still supporting OS/32, and the machines on which it ran, to some degree, according to the "Latest Facts" page on the Series 3200 on Concurrent's Web site.
(Some of my earliest systems programming work was at a summer job at Interdata, back in 1972; they were located near where my family lived. This was long before OS/32 existed - at the time I was there, some people there were working on an new OS that would add support for disk files that had names.)
Perhaps - although even better might have been a knowledge base (think HOWTO, if the term "knowledge base" sets off anti-Microsoft antibodies :-)) so that one wouldn't have to search the kernel source to find out what that message was saying, whether what it's saying could explain the problems one is seeing, and how to fix the problem.
It's nice that you can open the hood/bonnet and check the engine - sometimes the answer might not be in a knowledge base, and sometimes the answer in the knowledge base might not be the right answer for the particular problem you're seeing, and it's also nice that you can take a look and see what software is doing, see how it's doing it, and see why it's doing it, and it's nice that you can make it do something different - but it's not necessarily nice if you have to do that.
I'm not sure I'd go quite that far. Yes, it's all speculative, but one's position on what would've happened had Linux not been open-source may influence one's position on whether some project should be, or should become, open-source, and discussion might (or might not, of course...) help one eliminate speculations weakly supported or not supported by what data we have on what did happen with various software projects, and thus might (or might not...) help one make a better decision on said (hypothetical) project.
However, it may be wise to avoid coming up with Final Definite Conclusions about the effect of making projects open-source until there's a large set of projects that did, and didn't, go open-source, and enough time since they began to decide whether they succeeded or failed (bearing in mind, of course, Keynes' comment that, in the long run, we are all dead - or the statement it's claimed Zhou Enlai made when asked about the historical importance of the French Revolution, "It's too early to tell" - how long do you have to wait before you can decide what effect something had, and, by that time, is the answer meaningful or relevant?).
(And, of course, the projects may differ in ways having nothing to do with the openness of the source and that cause them to succeed or fail independently of whether they're open-source or not.)
I suppose System V Release 2 or so could also be thought of as the UNIX most commercial UNIXes these days derive from, either directly or indirectly, although SVR2's APIs, as noted, largely resemble a superset of V7, with some PWBisms, and some other additions, e.g. shared memory, and some changes, e.g. a different TTY driver "ioctl" interface.
SCO's only charging USD 100 for the license, it appears - and they also may "waive the license fee for a limited number of deserving applicants", where they suggest that "deserving" may include " student, unemployed, disabled, financially challenged, etc."
The sources are, of course, to PDP-11 UNIXes, so to actually run them you'd need a PDP-11, or a PDP-11 simulator.
The OS (or family of OSes) that has the biggest market share in the commercial and consumer marketplace has a command line. It's not as powerful a command line as that of UNIX, perhaps, and it may not be as necessary to drop down to the command line on those OSes as it is on at least some UNIX-flavored desktop OSes, but it does exist.
Give the PHB the perception that the command line doesn't exist, and they won't have the perception that there's anything odd to do on the command line.
I see no evidence to support that assertion. I suspect it'd be sufficient, at least at the "recompiling the kernel" level (and, mutatis mutandis, the "recompiling any other part of the system" level), to make it unnecessary to recompile the kernel for most purposes (e.g., making heavy use of loadable kernel modules, etc.).
(There will be applications ("embedded" systems, say, and perhaps at least some servers) where the ability to recompile the kernel may be a win, so, of course, the ability should't disappear.)
Perhaps there will be "end-user desktop" distributions where the source doesn't come on the distribution CD (the GPL doesn't require that what you distribute come with the source, you merely have to make the source available "for a charge no more than your cost of physically performing source distribution", "on a medium customarily used for software interchange"; that medium doesn't have to be the medium that contains the binary).
I'm not sure that's required to calm the fears of PHBs, however. I think the main reason for that might be to take one item off the list of options in the install program, to simplify the installation process, or, if the OS is pre-loaded on machines, it would be done to leave more disk space for pictures of the grandchildren or downloaded music or whatever.