The Ontario (Canada) system used to suffer
from problems in granting standing for
community (3-year) college courses to
prospective university students.
The technical solution was to
go through the standards for college
and university courses, and match
them at that level, so the university
can now say "Joe Student has taken COL-231
and COL-233, which matches out UNI-206
course".
Net results? The Universities are now
cooperating nicely with the Colleges.
Notably Seneca (College), which opened
their new campus at York (University)
Comment re Microsoft Office.DOC files in Microsoft Office 2003.
Jean Paoli, Senior Director XML Architecture for Microsoft has stated:
"We are acknowledging that end users who merely open and read government documents that are saved as Office XML files within software programs will not violate the license."
I and arguably my employer, a U.S. company doing business in Massachusetts, would almost certainly wish to write documents for the commonwealth in the.DOC format Massachusetts prefers. The Microsoft permission to merely read documents authored by the government of the Commonwealth seems less than responsive to an effort to allow non-monopoly software to be used by persons
wishing to communicate with Massachusetts.
I once worked for such a company, and
a colleague pointed out that they were
asking him to divest himself of all
the equity he had in previously
written programs.
All of a sudden, it started to look
expensive to insist on that clause, as
divestiture usually involves buying
out an interest. The discussion
got bumped up the the V.P. (Hi, Ian!)
who promptly struck the clause out.
My SiteSA changed the filesystem layout
one weekend, with several incorrigable users
(me!) logged on. All I noticed was slowness.
This was on Multics, which had to stay available
7/24, and for which there was support for changing
hardware (a brand new thesis on that one) and
softwware on the fly.
If they'd blown it during the
the first pack, they could revert
to a backup disk pack, fix the problem
and continue.
The method was continuous small changes,
applied in low-load periods, watched through
high-load priods and then accepted. After whic h
you applied the next small change to that
subsystem. You could be changing lots of
other subsystems at the same time, of course, so longas they weren't dependant on each other,a
and you could keep them straight in your head.
Have a look on google scholar for Multics and
continuous change.
These "trusted" computers are almost
the diametrical opposite of what the
"Orange Book" says must be done in
a Trusted Computer Security Evaluation:
Trusted Path? no, controlled by a 3rd party
Authentication? no, controlled by a 3rd party
Discretionary access control? yes
Mandatory access control? no, controlled by a 3rd party
Audit? only for the third-party
Labels? no
Label Integrity? no
Labelling on export/printing? yes
Assurance? no, controlled by a 3rd party
Covert Channels protection? no, built-in covert channels
All in all, enough to cerate a brand-new
level to as to the existing A1 through D".
Level "F", do not
buy for any reason (:-))
NutscrapeSucks writes
As soon as Java goes open source, I plan on forking it. Job 1 will be to add first class support for COM and XPCOM objects.
I hope that is an tease. If it's not, Microsoft will be really, really happy if you do.Perhaps you should apply to them for funding.
Your forked Java will only work well on Windows
with native COM, XPCOM and possibly NET. For your java to be "run everywhere", then
arguably you should contribute to the MONO
project.
By the way, the task of writing MONO
is only slightly smaller than writing WINE,
whose team has been working on it for
more than a few years (:-))
Because the current big bottleneck
is memory latency, either vendors
will add more cores and use the
memory bandwidth, or they'll
scale more and more poorly.
It makes good sense to fix
the bottleneck, because that's
where the problem lies. Improving
other parts which don't have problems,
according to Amdahl, is A Bad Idea (:-))
It's a fixation with the near future, from
people with little minds.
In this case, a
denial that the research department was smarter
than the marketing department, and an active
effort to harm them.
A parallel effort at H-P in particular was
to get rid of the Alpha and Precision
architectures, which threatened the
"common knowledge" that Intel was the
only successfully chip designer in the universe.
Another recent example, and one which might
well be related, was the "everyone will
run NT 4" fad, where a smart marketing department
(Microsoft's!) convinced a number of short-sighted
companies that there was no future outside of
WinTel. One of my favorite companies,
Mips (SGI) was killed by that one.
In fact, the marketing department works in the
present, and the research department in the
future. The wisest company I know (NBTel)
had them both reporting to the same person.
Net result? He had better connectivity in
his entire suburb than I have at work. And
he had it years earlier than I did in
Toronto.
So I voted with my feet, and went with
an employer who is isn't much respected here,
but who doesn't believe the unconscious lies of marketers
and stock analysis, looking for the
"quick profit" that never comes.
Discussed here last month, when Jimmy Wales
spoke about the problems of credibility.
See http://slashdot.org/article.pl?sid=05/01/07/224824 8&tid=149&tid=95
See the the IT Infrastructure Library site
at http://www.itil.org.uk
This is a set of books on different aspects of
IT Operations, and is widely used in the industry.
Of course, some people misuse it to create a
straight-jacket (MS, for one), and others use it to make a
sarong (Sun, SGI), but the basic cloth is there (:-))
Consider Amdahl's law: the compilation parallelizes well,
yielding.o's, followed by a big, slow, single-threaded link., which defines the
lower bound for how long the whole job takes.
Sun's a hardware company, why wouldn't they
support the software and chipsets that can
deliver good performance in the two- to four-way
market?
--dave (biased, you understand) c-b
cowboy76Spain writes: )
I do not understand how people can be so cinnic. They do believe municipal WiFi is free?
This is a very old debate, from at least 1789 in the U.S. The question in those days
was whether the federal government was justified
in spending money on free light signals for shipping.
The construction of a lighthouse, the staffing, maintenance and
supplies (whale oil initially, electricity later) was quite
expensive. Did the citizens imagine that the costs would not
be recovered from them in some other way...
The subject of the debate is traditionally called "universal goods",
and was discussed by John Stuart Mill in the 1844 "Essays on Some
Unsettled Questions of Political Economy". There's a relevant
quote from Book 5 Chapter 11 at
http://william-king.www.drexel.edu/top/eco/excerpt s/Mill.html
If you'd like to skip to the end of the debate, the common
resolution is for the local community to pay for the "natural
monopoly" component (in this case, the antenna farm) and allow
any business to use the service provided to offer values-added services
to customers. Any similarity to legal regiemes where one needs both an ISP and a cable
provider is not accidental (:-))
th poster said IT is commonly held as a cost center
Classic oops: if IT serves transactions for
sales, it's "part" of sales, a profit center.
As a capacity planner, I usually talk to
the business managers and say things like
Last christmas you rejected about 3% of your
purchase transactions. You say you're
growing by 6% per year, so next year you're going
to reject almost 9% of the christmas business.
The hardware to handle 10% more load will
cost you $X, and therefor, if the profit on
the average transaction is more than $X/(9% of the transaction count) you'll cover your costs
immediately.
Npt to mention avoiding pissing off the customers who you'd have rejected!
These are arguements to a profit center:
if you can credibly make them, they'll
dig up the money and force your boss to take it (:-))
A colleague suggests that 4 GHz may
be a hard frequency to exceed (in the
short run).
My leaky brain suggests that this
might correspond to the propogation
speed in silicon for a given path length
and a given process (eg, 90nm may give
us better results).
prostoalex writes: Windows Forms contains.NET framework classes for building GUI applications.
If you use classes which depend on.NET, your
application is dependent on code that's part
of a monopoly platform.
If you expect it to run anywhere than on
Windows you have to depend on MS not using
license terms, embrace-and-extend and patents
to make mono fail. Or they can just keep
changing the implementation as fast as they can
ship out updates, and wear the mono folks out
retaining the existing functionality, leaving
them unable to add to the framework.
Remember how long WINE took? And how
few apps ran under it for the first few years?
Indeed, how few run under it even now...
harikiri wrote:...even if MS lets them have insider info on say their filesystem, they can't release this info [...] because of NDA's and IP licensing restrictions.
I used to work with the Sun SMB team, who
had licensed code from AT&T, who in turn
licensed NT server code from Microsoft.
Only a small group of developers were permitted
access to the code, and they didn't even give fellow
Sun people doing rotations with them (me!) access.
Which was good, as I am a minor contributor
to Samba, and I didn't want to be polluted by
knowledge of other people's trade secrets.
This also applies to a lot of other stuff
licensed by Sun: I remember the effort it took to find
out what could be released and what couldn't
back when Sun first made Solaris 8 source available.
Slashdot Mirror
The technical solution was to go through the standards for college and university courses, and match them at that level, so the university can now say "Joe Student has taken COL-231 and COL-233, which matches out UNI-206 course".
Net results? The Universities are now cooperating nicely with the Colleges. Notably Seneca (College), which opened their new campus at York (University)
--dave
Jean Paoli, Senior Director XML Architecture for Microsoft has stated:
"We are acknowledging that end users who merely open and read government documents that are saved as Office XML files within software programs will not violate the license."
I and arguably my employer, a U.S. company doing business in Massachusetts, would almost certainly wish to write documents for the commonwealth in the .DOC format Massachusetts prefers. The Microsoft permission to merely read documents authored by the government of the Commonwealth seems less than responsive to an effort to allow non-monopoly software to be used by persons
wishing to communicate with Massachusetts.
--dave
All of a sudden, it started to look expensive to insist on that clause, as divestiture usually involves buying out an interest. The discussion got bumped up the the V.P. (Hi, Ian!) who promptly struck the clause out.
--dave
This was on Multics, which had to stay available 7/24, and for which there was support for changing hardware (a brand new thesis on that one) and softwware on the fly.
If they'd blown it during the the first pack, they could revert to a backup disk pack, fix the problem and continue.
The method was continuous small changes, applied in low-load periods, watched through high-load priods and then accepted. After whic h you applied the next small change to that subsystem. You could be changing lots of other subsystems at the same time, of course, so longas they weren't dependant on each other,a and you could keep them straight in your head.
Have a look on google scholar for Multics and continuous change.
--dave
Trusted Path? no, controlled by a 3rd party
Authentication? no, controlled by a 3rd party
Discretionary access control? yes
Mandatory access control? no, controlled by a 3rd party
Audit? only for the third-party
Labels? no
Label Integrity? no
Labelling on export/printing? yes
Assurance? no, controlled by a 3rd party
Covert Channels protection? no, built-in covert channels
All in all, enough to cerate a brand-new level to as to the existing A1 through D". Level "F", do not buy for any reason (:-))
--dave
I hope that is an tease. If it's not, Microsoft will be really, really happy if you do.Perhaps you should apply to them for funding.
Your forked Java will only work well on Windows with native COM, XPCOM and possibly NET. For your java to be "run everywhere", then arguably you should contribute to the MONO project.
By the way, the task of writing MONO is only slightly smaller than writing WINE, whose team has been working on it for more than a few years (:-))
It makes good sense to fix the bottleneck, because that's where the problem lies. Improving other parts which don't have problems, according to Amdahl, is A Bad Idea (:-))
In this case, a denial that the research department was smarter than the marketing department, and an active effort to harm them.
A parallel effort at H-P in particular was to get rid of the Alpha and Precision architectures, which threatened the "common knowledge" that Intel was the only successfully chip designer in the universe.
Another recent example, and one which might well be related, was the "everyone will run NT 4" fad, where a smart marketing department (Microsoft's!) convinced a number of short-sighted companies that there was no future outside of WinTel. One of my favorite companies, Mips (SGI) was killed by that one.
In fact, the marketing department works in the present, and the research department in the future. The wisest company I know (NBTel) had them both reporting to the same person. Net result? He had better connectivity in his entire suburb than I have at work. And he had it years earlier than I did in Toronto.
So I voted with my feet, and went with an employer who is isn't much respected here, but who doesn't believe the unconscious lies of marketers and stock analysis, looking for the "quick profit" that never comes.
--dave
--dave
This is a set of books on different aspects of IT Operations, and is widely used in the industry. Of course, some people misuse it to create a straight-jacket (MS, for one), and others use it to make a sarong (Sun, SGI), but the basic cloth is there (:-))
It's orthogonal to the CMU maturity model.
-dave
--dave
--dave
Sun's a hardware company, why wouldn't they support the software and chipsets that can deliver good performance in the two- to four-way market? --dave (biased, you understand) c-b
This is a very old debate, from at least 1789 in the U.S. The question in those days was whether the federal government was justified in spending money on free light signals for shipping. The construction of a lighthouse, the staffing, maintenance and supplies (whale oil initially, electricity later) was quite expensive. Did the citizens imagine that the costs would not be recovered from them in some other way...
The subject of the debate is traditionally called "universal goods", and was discussed by John Stuart Mill in the 1844 "Essays on Some Unsettled Questions of Political Economy". There's a relevant quote from Book 5 Chapter 11 at http://william-king.www.drexel.edu/top/eco/excerpt s/Mill.html
If you'd like to skip to the end of the debate, the common resolution is for the local community to pay for the "natural monopoly" component (in this case, the antenna farm) and allow any business to use the service provided to offer values-added services to customers. Any similarity to legal regiemes where one needs both an ISP and a cable provider is not accidental (:-))
--dave
Almost everything applies to sceens, too.
--dave
Don't care: it falls under the "Foreign Corrupt Practices Act: (as noted on Groklaw).
--dave
Classic oops: if IT serves transactions for sales, it's "part" of sales, a profit center.
As a capacity planner, I usually talk to the business managers and say things like
These are arguements to a profit center: if you can credibly make them, they'll dig up the money and force your boss to take it (:-))
--dave
Instead of one a third-party proprietary chip that's failed, resurrect one of the two RISC designs H-P dropped in favor of the bad one.
--dave
My leaky brain suggests that this might correspond to the propogation speed in silicon for a given path length and a given process (eg, 90nm may give us better results).
--dave
If you use classes which depend on .NET, your
application is dependent on code that's part
of a monopoly platform.
If you expect it to run anywhere than on Windows you have to depend on MS not using license terms, embrace-and-extend and patents to make mono fail. Or they can just keep changing the implementation as fast as they can ship out updates, and wear the mono folks out retaining the existing functionality, leaving them unable to add to the framework.
Remember how long WINE took? And how few apps ran under it for the first few years? Indeed, how few run under it even now...
--dave
--dave
I used to work with the Sun SMB team, who had licensed code from AT&T, who in turn licensed NT server code from Microsoft. Only a small group of developers were permitted access to the code, and they didn't even give fellow Sun people doing rotations with them (me!) access.
Which was good, as I am a minor contributor to Samba, and I didn't want to be polluted by knowledge of other people's trade secrets.
This also applies to a lot of other stuff licensed by Sun: I remember the effort it took to find out what could be released and what couldn't back when Sun first made Solaris 8 source available.
--dave
--dave
Mind you, Plan 9 From Outer Space is seriously retro!
--dave