What you say is: for internal use, Open Source is just like any license, except it comes with a source license. Usually we have to pay a lot for a source license, so that's a really good deal.
Generally when we have a source license and fix a problem or add some functionality we need, we send our patch to the vendor and ask them to include it in future releases. This is to our benefit because it reduces our maintence load since we don't have to re-patch each new release ourselves. The same logic continues to apply with an "open source" source license.
For inclusion of the software in a product we ship, particularly if we have customized the software, we need to look at the specific license. In general, in looking at the license and deciding whether to use the software, "community good will" should enter into our calculations in the open source case, while that would not usually be a consideration with proprietary software.
Using "open source" in shipping products, and handling it well, is an opportunity for advertizing and marketing. It can constitute a competitive advantage in the marketplace, because many customers prefer software they are already familiar with, which they are confident in, and which they can easily examine and modify.
That's not fair; many sellers engage in price discrimination, i.e. they sell essentially the same product to different groups based on each group's ability to pay. This is particularly common when marginal costs are low but sunk costs are high. DVDs are a good example, with "region codes" serving that sole purpose. University education is another nice one, with "scholarships" used to set radically different price points for different customers. Clothes are another classic example. Identical cars are often sold at dramatically different prices by changing nothing but the logo on the hood. The most familiar example is probably airplane seats; the airlines have raised price discrimination to a high art.
(This doesn't change the fact that Microsoft often engages in unethical business practices. Just not in this case, at least until them have the whole country locked in...)
Good one. You can skew the statistics even more with the trick actually used by the airlines in their safety claims: "fatality-producing incidents per million passenger miles." Then you'll be down to 0.001.
Unlike the other characters in the LOTR movies, including other CGI ones, Gollum was terrible. He practically chewed the scenery he was overacting so hard. Worse than Kirk! And he did all kinds of physically implausible things, like running full-tilt and being stopped cold by a noose around his neck and just picking himself up off the rock he landed on his head on none the worse for wear. He wasn't even bruised - not a scratch. Not to mention his conspicuously un-broken neck.
Basically, it was bugs-bunny animation physics.
In other words: Gollum didn't get a "best supporting actor" nomination because he didn't deserve one: he sucked.
If you are acting as a professional then you have certain obligations. These are summed up in various "codes of ethics for professional engineers" documents, such as the National Society for Professional Engineers "Code of Ethics for Engineers" at National Society for Professional Engineers' Code of Ethics for Engineers,
which states in part:
I. Fundamental Canons
Engineers, in the fulfillment of their professional duties, shall:
1. Hold paramount the safety, health and welfare of the public.
2. Perform services only in areas of their competence.
3. Issue public statements only in an objective and truthful manner.
4. Act for each employer or client as faithful agents or trustees.
5. Avoid deceptive acts.
6. Conduct themselves honorably, responsibly, ethically, and lawfully so as to enhance the honor, reputation, and usefulness of the profession.
I call particular attention to point 4, and would emphasize that your clients deserve your honest professional opinion concerning matters in your area of competence, the same as they would deserve if you were an employee with accompanying fiduciary responsibilities. This is what it means to act as their agent. Under certain circumstances, not calling their attention to something you know to be a problem could even be a deceptive act, and you certainly would not want to do that!
The CRIS CPU architecture seems reasonable for an embedded CPU
optimized for code volume, but to call it RISC is outrageous.
There is nothing RISCy about CRIS. In fact it is pretty much a
classic CISC, highly reminiscent of the VAX.
Look at the complex addressing modes and variable-length
instructions: hallmarks of a CISC. To quote chapter
2 of the documentation,
2.3 DATA ORGANIZATION IN MEMORY...
Data can be aligned to any address. If the data crosses a 32-bit
boundary, the CPU will split the data access into two separate
accesses. The use of unaligned word and dword data will thus
degrade the performance.
...
2.4.1 Addressing Modes
The CRIS CPU has four basic addressing modes. These modes are
encoded in the mode field of the instruction word. The basic
addressing modes are:
Quick immediate mode
Register mode
Indirect mode
Autoincrement mode (with immediate mode as a special case)
More complex addressing modes can be achieved by combining the
basic instruction word with an addressing mode prefix word. The
complex addressing modes are:
Indexed
Indexed with assign
Offset
Offset with assign
Double indirect
Absolute
The addressing modes of the CRIS CPU are
Assembly syntax; Addressing mode i, j Quick immediate Rn Register Pn Special register [Rn] Indirect [Rn+] Post increment x, u Byte immediate xx, uu Word immediate xxxx, uuuu Dword immediate [Rn+Rm.m] Indexed [Rp=Rn+Rm.m] Indexed with assign [Rn+[Rm].m] Indirect offset [Rn+[Rm+].m] Autoincrement offset [Rn+x] Immediate byte offset [Rn+xx] Immediate word offset [Rn+xxxx] Immediate dword offset [Rp=Rn+[Rm].m] Indirect offset with assign [Rp=Rn+[Rm+].m] Autoincrement offset with assign [Rp=Rn+x] Immediate byte offset with assign [Rp=Rn+xx] Immediate word offset with assign [Rp=Rn+xxxx] Immediate dword offset with assign [[Rn]] Double indirect [[Rn+]] Double indirect with auto increment [uuuu] Absolute
curmudgeon take on libraries and technology
on
Libraries Are 31337
·
· Score: 1
Curmudgeon take on libraries and technology: http://www.cs.unm.edu/~bap/library-in fo-commons.ht ml
Permission to use, copy, or modify this software and its documentation
for
education and research purposes only...
For any other uses of this
software, in original or modified form, including but not limited to
distribution in whole or in part, specific prior permission must be
obtained from MIT.
This means that if little Johnny uses starlogo at school, so he borrows a CD and installs it on his computer at home so he can play with it there too, and then he writes a starlogo program to help daddy manage his business' inventory, they are in violation.
As an educator, that is too risky for me. I'd rather give students stuff they can use however they want, rather than forbidding use for non-class purposes or sending them home with a CDROM only after they've heard a long list of warnings.
Date: Fri, 14 Jan 1994 04:29:08 -0500 Cc: unix-haters@mc.lcs.mit.edu Subject: Re: Re: Re- Stalin never had it so good...
I think we can all agree that the odious typed smiley faces so beloved of the mass of ignorant unix nethead weenies, who never learned to express themselves in written form, and who therefore clutch at these pathetic glyphs as if by sprinkling a few grains of malodorous typographical powder upon a heavyhanded ignorant misspelled missive it will magically become imbued with lighthearted humor and wise witty sarcasm, have no place on unix-haters.:)
My lab at the University of New Mexico has a project in BCI, in collaboration with a group at Sarnoff Labs in Princeton. We are using novel signal processing methods and new kinds of brain imaging instrumentation and hope to get higher bit rates and lower cognitive loads than current systems. We're always looking for good people to work on this kind of stuff, especially as grad students or postdocs.
You are treating the GPL's requirement to distribute sources as a liability, as something that may cost your company money. In fact, it is actually a revenue generation opportunity!
Your physical mailbox at home is full of CDROMs sent to you for free. Apparently it is economic to send CDROMs to non-computer people in the hope that they will use them. It is therefore undoubtedly the case that companies would line up around the block for a list of *developers* who have *asked* for a CDROM, ie professional computer people who purchase products, who have money, and who are guaranteed to actually use the CDROM they are sent as something other than a coaster.
You could make the process of asking for a copy of the sources very easy, just a matter of either filling out a form on the web, calling an 800 number, or sending a postcard. Maybe require $5 (using a credit card order form on the web) just as a proof of seriousness, or maybe not. Then you take this list, and you SELL it to a "freebie" computer magazine targetted at developers. Or to a computer catalog company. Or to a Linux software vendor. Or whatever. They send everyone their magazine or catalog or whatever, along with a CDROM containing (a) your sources (ie what the person actually asked for), plus (b) anything else they want to put on that disk, eg demo products, advertisements, a copy of their linux distribution, a copy of their catalog, whatever.
If you cannot find a dozen companies that would pay oodles for this privilege, then your company needs a new marketing dept.
The Bayesian community has known about this for many years. It is a log Gaussian, which is the prior commonly used for SCALE PARAMETERS in Bayesian estimation. It is interesting that it applies to other scale parameters, but it's what you'd expect, not some big breakthrough.
This modality - MRI - gives excellent spatial resolution. Unfortunately, it is not so good with temporal information, which can be at best on the order of a second, which is much slower than the brain processes information. In my lab we are using magnetoencepalography and related methods, in concert with MRI, to look at the temporal dynamics of the brain.
In other words, MRI gives a schematic of the brain, and fMRI tells you which parts get warm when doing certain tasks. We're trying to use MEG like a logic probe, to look at timing.
Our code base, both for signal processing and for visualization, is all developed on Debian GNU/Linux machines (both i386 and Alpha) and will all to be released under the GLP. It is also all being ported to SGIs and to large Linux clusters.
If you're interested in figuring out how the brain works and want to get a PhD or MS in CS at a really funky department while hacking Linux and playing with gonzo brain imaging data, don't be shy - get in touch.
Slashdot Mirror
Generally when we have a source license and fix a problem or add some functionality we need, we send our patch to the vendor and ask them to include it in future releases. This is to our benefit because it reduces our maintence load since we don't have to re-patch each new release ourselves. The same logic continues to apply with an "open source" source license.
For inclusion of the software in a product we ship, particularly if we have customized the software, we need to look at the specific license. In general, in looking at the license and deciding whether to use the software, "community good will" should enter into our calculations in the open source case, while that would not usually be a consideration with proprietary software.
Using "open source" in shipping products, and handling it well, is an opportunity for advertizing and marketing. It can constitute a competitive advantage in the marketplace, because many customers prefer software they are already familiar with, which they are confident in, and which they can easily examine and modify.
(This doesn't change the fact that Microsoft often engages in unethical business practices. Just not in this case, at least until them have the whole country locked in...)
Anyway isn't the shuttle just going in circles?
Basically, it was bugs-bunny animation physics.
In other words: Gollum didn't get a "best supporting actor" nomination because he didn't deserve one: he sucked.
Look at the complex addressing modes and variable-length instructions: hallmarks of a CISC. To quote chapter 2 of the documentation,
Curmudgeon take on libraries and technology:n fo-commons.ht ml
http://www.cs.unm.edu/~bap/library-i
As an educator, that is too risky for me. I'd rather give students stuff they can use however they want, rather than forbidding use for non-class purposes or sending them home with a CDROM only after they've heard a long list of warnings.
Date: Fri, 14 Jan 1994 04:29:08 -0500
:)
Cc: unix-haters@mc.lcs.mit.edu
Subject: Re: Re: Re- Stalin never had it so good...
I think we can all agree that the odious typed smiley faces so beloved
of the mass of ignorant unix nethead weenies, who never learned to
express themselves in written form, and who therefore clutch at these
pathetic glyphs as if by sprinkling a few grains of malodorous
typographical powder upon a heavyhanded ignorant misspelled missive it
will magically become imbued with lighthearted humor and wise witty
sarcasm, have no place on unix-haters.
Can the GNU toolset target these gaming devices, and is there an easy way (ie short of burning a ROM) to get code into them?
Academic research of many sorts requires such access:
- research into compression algorithms
- computer vision research, eg face recognition
- medical research (looking at tiny skin features, tracking eye motion, muscle tremmors)
- social science research that needs to view tiny objects in the background, such as advertisements and product labels
- tracking actors through their careers via face recognition
- meteorological studies looking at cloud features, night skies
- history of science and technology, optics & camera technology
- ecology, looking at foliage, insects, birds
There are all potential fair uses. Some of them are beyond our current technical capabilities. But not for long.My lab at the University of New Mexico has a project in BCI, in collaboration with a group at Sarnoff Labs in Princeton. We are using novel signal processing methods and new kinds of brain imaging instrumentation and hope to get higher bit rates and lower cognitive loads than current systems. We're always looking for good people to work on this kind of stuff, especially as grad students or postdocs.
Your physical mailbox at home is full of CDROMs sent to you for free. Apparently it is economic to send CDROMs to non-computer people in the hope that they will use them. It is therefore undoubtedly the case that companies would line up around the block for a list of *developers* who have *asked* for a CDROM, ie professional computer people who purchase products, who have money, and who are guaranteed to actually use the CDROM they are sent as something other than a coaster.
You could make the process of asking for a copy of the sources very easy, just a matter of either filling out a form on the web, calling an 800 number, or sending a postcard. Maybe require $5 (using a credit card order form on the web) just as a proof of seriousness, or maybe not. Then you take this list, and you SELL it to a "freebie" computer magazine targetted at developers. Or to a computer catalog company. Or to a Linux software vendor. Or whatever. They send everyone their magazine or catalog or whatever, along with a CDROM containing (a) your sources (ie what the person actually asked for), plus (b) anything else they want to put on that disk, eg demo products, advertisements, a copy of their linux distribution, a copy of their catalog, whatever.
If you cannot find a dozen companies that would pay oodles for this privilege, then your company needs a new marketing dept.
The Bayesian community has known about this for many years. It is a log Gaussian, which is the prior commonly used for SCALE PARAMETERS in Bayesian estimation. It is interesting that it applies to other scale parameters, but it's what you'd expect, not some big breakthrough.
This modality - MRI - gives excellent spatial resolution. Unfortunately, it is not so good with temporal information, which can be at best on the order of a second, which is much slower than the brain processes information. In my lab we are using magnetoencepalography and related methods, in concert with MRI, to look at the temporal dynamics of the brain.
In other words, MRI gives a schematic of the brain, and fMRI tells you which parts get warm when doing certain tasks. We're trying to use MEG like a logic probe, to look at timing.
Our code base, both for signal processing and for visualization, is all developed on Debian GNU/Linux machines (both i386 and Alpha) and will all to be released under the GLP. It is also all being ported to SGIs and to large Linux clusters.
If you're interested in figuring out how the brain works and want to get a PhD or MS in CS at a really funky department while hacking Linux and playing with gonzo brain imaging data, don't be shy - get in touch.
"We are desparate
Get used to it
We are desparate
Get used to it"
-- X
(The window system or the rock
band? You decide.)