Microsoft was already months into a massive project aimed at taking down Google when the truth began to dawn on Bill Gates. It was December 2003. He was poking around on the Google company website and came across a help-wanted page with descriptions of all the open jobs at Google. Why, he wondered, were the qualifications for so many of them identical to Microsoft job specs...
Adult male consumes about 2000 kilocalories a day.
WAY low: maybe 2500 for 5'8" "sedentary" (depends a lot on body size;
lots more for 6'5" than 5'5" -- law of cubes),
more for "active", and a lot more for very active.
When I was in graduate school, and an active fencing competitor, summers were "endurance training" time. I'd eat 6000+ calories a day,
and wind up the summer 15 pounds lighter in
September than I was in May (some of that loss will have been leg-muscle bulk, however).
After reading the article, I realised that the frontside bus was shared. I didn't expect that. It seems to be a transitory solution in order to have the "first dual-core" CPUs on the market. When AMD releases theirs I expect them to have a superior solution.
AMD64 has had the circuitry for dual-core on-chip
memory controllers from the very first -- they just didn't have the second CPU core. For a good discussion of the differences, see
http://www.linuxhardware.org/features/05/04/21/174 7217.shtml
at LinuxHardware.
(1) a work prepared by an employee within the scope of his or her employment; or
(2) a work specially ordered or commissioned for use as a contribution to a collective work, as a part of a motion picture or other audiovisual work, as a translation, as a supplementary work, as a compilation, as an instructional text, as a test, as answer material for a test, or as an atlas, if the parties expressly agree in a written instrument signed by them that the work shall be considered a work made for hire. For the purpose of the foregoing sentence, a "supplementary work" is a work prepared for publication as a secondary adjunct to a work by another author for the purpose of introducing, concluding, illustrating, explaining, revising, commenting upon, or assisting in the use of the other work, such as forewords, afterwords, pictorial illustrations, maps, charts, tables, editorial notes, musical arrangements, answer material for tests, bibliographies, appendixes, and indexes, and an "instructional text" is a literary, pictorial, or graphic work prepared for publication and with the purpose of use in systematic instructional activities.
If your work was not in the course of your normal
duties, and if it was not *specifically* commissioned by your employer, then as I read it, Federal law
says the work is yours, unless you executed a proper
transfer of ownership.
A transfer of copyright ownership, other than by operation of law, is not valid unless an instrument of conveyance, or a note or memorandum of the transfer, is in writing and signed by the owner of the rights conveyed or such owner's duly authorized agent.
says that the transfer must have been specific
to the item in question, and in writing.
I think you should hit them up for extortion...
and maybe hit *them* for copyright violation
if they try to hang onto the stuff.
Note that the copyright transfier stuff has come up in SCO vs.Novell, recently -- see GrokLaw.
Complete and general aliasing analysis is NP-complete, IIRC. OTOH, it is somewhat easier in Fortran to stay within a linear-time "guaranteed
not to alias" subset of programming constructions...
My complaint is that until F03 comes out fully supported by the vendors, struct/TYPEs aren't quite first class citizens in Fortran, as exemplified by the
following
It is *possible* to write C that runs as fast as Fortran for heavy math. However, it involves hand-optimizing your C until this happens.
It is a matter of record that during the early phases of gcc and development,
Craig Burley had an enormous problem convinding
Richard Stallman that a family of optimizations
associated with "true" arrays were worthwhile,
and in fact necessary for the back end of the system. Once they were put in, performance was approximately doubled for a broad range of Fortran codes.
And you're telling me that all this hand-optimization can be done by a routine "good"
programmer, when one of the premier C programmers
of the age couldn't see its value ?!
For what it's worth, the first automobile to exceed 200 MPH was a Stanley Steamer, outfitted with an
aerodynamic body (IIRC, the year was 1919,
the location Daytona Beach).
Unfortunately, they didn't understand aerodynamics
as well then as we do now, and on the next run the
car hit a bump and became the first automobile to
fly more than 500 feet -- totalling the car and
killing the driver.
AFAIK, Word 97-2003 have the same file format. Excepting some possible formatting issues, reading the documents shouldn't be a problem...
Can you say, clueless!?
There are incompatibilities between the paragraph and character styles and the numbering mechanisms among the
versions of Word you talk about (97/2000/XP), and
going back and forth among them is a sure way to
almost-irremediable document corruption. As a
corporate-law attorney, my wife runs into this
problem all the time.
Word can't deal with it; the commercial
product for cleaning up the mess runs $5000/seat
and many law firms consider it well worth the
price. (Or you can use the industrial-strength.doc-parser found in abiword or
OpenOffice.org:-).)
The PathScale folks (who started out as a spin-off from Cray's compiler group) worked extensively with AMD to construct "state-of-the-art" C and Fortran
compilers for AMD64. See
http://www.pathscale.com/index.html
Since the code for these
benchmarks is available, it would have been really
interesting (for me--as a developer/environmental
modeler who compiles his own codes) to see what
performance boost these compilers would have given
(as compared with default "gcc" builds)...
A lot more work, I'll admit.
Windows scrollbars and widgets that drive me batty:
MS forces me to manually snoopervise scroll operations, instead of just "grab the scroll tab and yank it down." If your mouse strays outside that narrow little scrollbar rectangle, you lose...
Windows focus policies (click-to-focus only)
drive me nuts!
Applications that claim to know my job better than I do, but blow it! I hate uppity software!
(e.g., when I write "netCDF" I have in fact used exactly the capitalization I intended. I hate apps that insist on changing it
to something that is wrong.
If I couldn't do a better job of software design than that, I'd get out of the business!
I had one reaction to your recent article about Linux and other
open source software:
What did de Tocqureville himself say about voluntary community
associations in America, how they were _uniquely American_, and
uniquely effective because of the way they got out from under
the rigid guild-and-government strictures of the Old World?
And how are the current IP-law trends different from an attempt
to reimpose those guild-and-government strictures on voluntary
community associations that now (thanks to the Internet) can span
the world?
and, generally, in reaction to trends in copyright and other IP law over
the past century, a further question:
How is the present unlimited copyright term consistent with the
demands of the Constutution? (Given that (1) when copyright term
may be extended at will by Congress, that term is by definition
NOT limited in mathematical terms (note that I _am_ a professional
mathematician (Ph.D.1978, MIT)); (2) when there is no experiment
that I can perform which distinguishes current copyright term
from unlimited term, then that term is operationally unlimited;
and (3) when no work I can find has had its copyright term expire
during my lifetime, _nor can I expect such an expiration to happen
during the actuarial remainder thereof_, then copyright term is
not limited in human terms.)
Absent a satisfactory answer on these mattters, I can only conclude
that your claim to follow de Tocqueville is hypocritical at best.
Better would be to have a choice between menu systems that the user can make easily fromthe
control panel: something like the following
set of choices would be appropriate:
As soon as I see the page, I see the demand to
present it in a font significantly smaller than
I have told my browser I want to see ("...font-size : 90%;") to be both obnoxious and arrogant. This
is made the worse by the tendency of some CSS hackers (not you, in this case) to make the font
much smaller than my "never see" setting.
Mozilla's "zoom" feature helps this somewhat, but it still makes pages made with this CSS more trouble to read than they would be otherwise.
It is a pain in the ass to have to edit
the CSS before you can read a page comfortably. [Thanks, Jesse, for the
bookmarklets to do that,
http://www.squarefree.com/bookmarklets/]
A
manifold is a generalization of a surface (on a surface like a torus, you can move in two independent directions at any point; on an n dimensional manifold, you can move in n independent directions. Space-time is a 4-manifold.) Manifolds are the most general sorts of objects you can write differential equations and integrals on.
Elliptic differential equations (very informally speaking) are differential equations that act like the equations for equilibrium problems.
THEOREM Elliptic
differential equation systems have finite dimensional solution sets (Hodge, Fredholm).
(That dimensionality is an integer)
THEOREM
That dimensionality is a topological invariant of the manifold. (de Rham).
The Heat Equation solution technique
for elliptic differential equations leads to the computation of an integral over
the manifold (not sure the best reference here,
probably Friz John from NYU-Courant). (The result of that integral is a real number.)
Theorem (Atiyah-Singer) The (real-number) integral coming from the Heat Equation solution technique is the same as the (integer) topological invariant coming from the dimensionality of the solution space.
This says that the topology ("how many holes in the torus?") is intimately tied up with the solvability of differential equations (an entirely
different branch of math); moreover, the differential equations (as occur in mathematical
physics) have solution properties that generate
integers (tying in to quantum mechanics).
The only problems I have with them come from boneheaded websites that check the browser and then refuse to allow any none IE browser to access the site. How clueless is that.
Not as clueless as the ones that claim to do such
a check, and then reject you for not having the very browser that in fact you are
using. (They claimed to support IE and
Mozilla, then rejected Mozilla 1.6 (Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040113)
I ran into one of those yesterday on a sporting
goods sales site... wrote them a nastygram quoting
their rejection-page back to them, together with
my browser identity, then
asking whether I should expect the same kind
of bullshit from their merchandise that I find
in their web site design.
The article talks about spinning the coin around
a horizontal axis as being the
least-biased way to flip a coin, slanted axes
having biases.
An interesting alternative is to flip the coin
so that it lands on a smooth floor, spinning on
a vertical axis. Then the uneven
distribution of mass between the head-side and the
tail-side will cause a bias.
It is my experience that dimes and quarters are
nearly unbiased for this test, whereas nickels are
heavily biased (pun intended) toward
tails. [In a past life, I taught a
statistics class for which I assigned
daily homework, deciding whether or not to take
it up on the basis of a coin flip at the end of
class.
On days for which I really didn't want to
spend all evening grading papers, I would use a
nickel; I'd use a much-fairer quarter on other
days. And none of the class caught on... ]
The Intel x87 uses Pade representations in hardware,
accurate to IEEE REAL*10 EXTENDED, i.e., about 24
significant digits.
Otoh, the P4 SSE2 uses a vectorized software model
that doesn't have these; I don't know whether the MS compiler generates x87 hardware code or SSE2
vectorized software.
Java specifies using a 32-bit
model for these functions, and is probably doing them in software. But what software? And does it use the vectorized SSE2?
IANAL, but the trouble is this: They claim it's not a sale.
What would be very helpful is if you
could establish the "looks like a duck, quacks
like a duck" precedent that as a matter of public policy it is indistinguishable from a sale, so that your comment would apply.
Yeah, but Los Alamos Computers (http://www.laclinux.com/en/Linux_Computers),builder of ESR's "Ultimate Linux Box" will build you a
nicely fitted out dual-Opteron system (8GB RAM, dual 160GB SATA disks,...) for about $2900...
Slashdot Mirror
The computer-architecture blog "realworldtech" has been slogging this one out in extreme detail (particularly as regards integer performance, where the Itanium has seemed to lag); see http://www.realworldtech.com/forums/index.cfm?acti on=detail&PostNum=3510&Thread=1&entryID=52549&room ID=11
When I was in graduate school, and an active fencing competitor, summers were "endurance training" time. I'd eat 6000+ calories a day, and wind up the summer 15 pounds lighter in September than I was in May (some of that loss will have been leg-muscle bulk, however).
For benchmarks relating to serious DB and web use, see this review by Anand Shempi: http://www.anandtech.com/cpuchipsets/showdoc.aspx? i=2397
or these two at FiringSquad:
http://www.firingsquad.com/hardware/amd_dual-core_ opteron_875/
and http://www.firingsquad.com/hardware/colfax_dual_op teron/
Then, (17 USC 204, q.v.http://www.law.cornell.edu/uscode/html/uscode1 7/usc_sec_17_00000204----000-.html
says that the transfer must have been specific to the item in question, and in writing.I think you should hit them up for extortion... and maybe hit *them* for copyright violation if they try to hang onto the stuff.
Note that the copyright transfier stuff has come up in SCO vs.Novell, recently -- see GrokLaw.
Complete and general aliasing analysis is NP-complete, IIRC. OTOH, it is somewhat easier in Fortran to stay within a linear-time "guaranteed not to alias" subset of programming constructions...
My complaint is that until F03 comes out fully supported by the vendors, struct/TYPEs aren't quite first class citizens in Fortran, as exemplified by the following
fwiw
And you're telling me that all this hand-optimization can be done by a routine "good" programmer, when one of the premier C programmers of the age couldn't see its value ?!
Unfortunately, they didn't understand aerodynamics as well then as we do now, and on the next run the car hit a bump and became the first automobile to fly more than 500 feet -- totalling the car and killing the driver.
There are incompatibilities between the paragraph and character styles and the numbering mechanisms among the versions of Word you talk about (97/2000/XP), and going back and forth among them is a sure way to almost-irremediable document corruption. As a corporate-law attorney, my wife runs into this problem all the time.
Word can't deal with it; the commercial product for cleaning up the mess runs $5000/seat and many law firms consider it well worth the price. (Or you can use the industrial-strength .doc-parser found in abiword or
OpenOffice.org:-) .)
Since the code for these benchmarks is available, it would have been really interesting (for me--as a developer/environmental modeler who compiles his own codes) to see what performance boost these compilers would have given (as compared with default "gcc" builds)... A lot more work, I'll admit.
- Windows scrollbars and widgets that drive me batty:
MS forces me to manually snoopervise scroll operations, instead of just "grab the scroll tab and yank it down." If your mouse strays outside that narrow little scrollbar rectangle, you lose...
- Windows focus policies (click-to-focus only)
drive me nuts!
- Applications that claim to know my job better than I do, but blow it! I hate uppity software!
(e.g., when I write "netCDF" I have in fact used exactly the capitalization I intended. I hate apps that insist on changing it
to something that is wrong.
If I couldn't do a better job of software design than that, I'd get out of the business!Dear Dr. Brown:
I had one reaction to your recent article about Linux and other open source software:
and, generally, in reaction to trends in copyright and other IP law over the past century, a further question: Absent a satisfactory answer on these mattters, I can only conclude that your claim to follow de Tocqueville is hypocritical at best.Sincerely--
Carlie J. Coats, Jr., Ph.D.
Mozilla's "zoom" feature helps this somewhat, but it still makes pages made with this CSS more trouble to read than they would be otherwise. It is a pain in the ass to have to edit the CSS before you can read a page comfortably. [Thanks, Jesse, for the bookmarklets to do that, http://www.squarefree.com/bookmarklets/]
I ran into one of those yesterday on a sporting goods sales site... wrote them a nastygram quoting their rejection-page back to them, together with my browser identity, then asking whether I should expect the same kind of bullshit from their merchandise that I find in their web site design.
idiot bastards!
An interesting alternative is to flip the coin so that it lands on a smooth floor, spinning on a vertical axis. Then the uneven distribution of mass between the head-side and the tail-side will cause a bias.
It is my experience that dimes and quarters are nearly unbiased for this test, whereas nickels are heavily biased (pun intended) toward tails . [In a past life, I taught a statistics class for which I assigned daily homework, deciding whether or not to take it up on the basis of a coin flip at the end of class. On days for which I really didn't want to spend all evening grading papers, I would use a nickel; I'd use a much-fairer quarter on other days. And none of the class caught on... ]
It even allows you to prepare a boot CD for one machine from a backup-set, hosted by another...
Otoh, the P4 SSE2 uses a vectorized software model that doesn't have these; I don't know whether the MS compiler generates x87 hardware code or SSE2 vectorized software.
Java specifies using a 32-bit model for these functions, and is probably doing them in software. But what software? And does it use the vectorized SSE2?
What would be very helpful is if you could establish the "looks like a duck, quacks like a duck" precedent that as a matter of public policy it is indistinguishable from a sale, so that your comment would apply.
Yeah, but Los Alamos Computers (http://www.laclinux.com/en/Linux_Computers) ,builder of ESR's "Ultimate Linux Box" will build you a
nicely fitted out dual-Opteron system (8GB RAM, dual 160GB SATA disks, ...) for about $2900...