I taught myself Fortran 40 years ago when I was 16 (long story about how that happened that's not relevant here). I would have been so much better off if there had been someone looking over my shoulder from time to time and making suggestions about style. I estimate that it took two years of computer science in college before I unlearned all of the bad habits I picked up in three months...
I recall a question used (many years ago) on some OCS exams. A detailed list of available materials was supplied, along with a sketch of the terrain surrounding a portion of a stream. The question was "How would you build a bridge capable of carrying jeeps across this stream?" The correct answer was "Sergeant, take these men and this pile of stuff and build a bridge across this stream. I'll be back in three hours." Some incorrect answers did get people into various specialist training programs.
More than 50,000,000 annually. Automotive World reported that global commercial vehicle production for 2008 was 70.5 million units, including 52.6 million passenger cars, 13.6 million light commercial vehicles, and 3.6 million trucks. Total production was down 10.8% from 2007 due to the effects of the world-wide recession. Production in China and India for local sales are increasing very rapidly.
I read somewhere, many years ago, that sci-fi is popular in good times, when people in general are looking forward to the future, and fantasy is popular in bad times when people are afraid of the future.
Indeed.
For many people, the worst parts of previous generations' speculative fiction
appears to be coming true.
The giant corporations are winning.
Ask people if they think it more likely that
genetic research will result in exciting new medical treatments or
be used by enormous health insurance companies to deny coverage.
The Luddites are winning.
Polls show that almost as many Americans
believe in creationism as evolution.
I find it disturbing that "If This Goes On--"
could be Heinlein's most accurate social forecast.
The problems keep turning out to be harder than most people thought.
LEO is still a bloody expensive place to get to.
Commercial nuclear fusion is always 30 years away.
We'll probably never get flying cars.
The general attitude towards engineering seems to have changed.
We went from the neutron as a theoretical particle to 100 commercial reactors
in 50 years;
but nuclear waste is regarded is a problem that engineers won't solve
even if given hundreds of years.
The Club of Rome's forecasts are turning out to be
depressingly accurate.
Many economists now believe that
the Baby Boomers' kids will be
the first generation in the US with a lower standard of living than their parents.
I spent the last three years on the budget staff for a state legislature. I feel quite comfortable saying that it is almost impossible, politically, for states to accumulate a meaningful rainy-day fund in the contemporary political climate -- for a medium-sized state, say a billion or more. In my medium-sized state, a billion would have been about half as much as was/is needed to fill the "revenue gap" this recession. Once the surplus starts to pile up, there is enormous pressure to cut tax rates rather than grow the surplus to the appropriate size. We have an easy citizen initiative process; if the state accumulated a billion dollar reserve I can guarantee there would be a ballot initiative to refund much of that to taxpayers and the initiative would have a very good chance of passing. Polls show consistently that our voters believe they can have both high spending and low taxes and are frustrated that the elected officials don't deliver it.
An exception where states do manage to accumulate reasonable surpluses is in their unemployment insurance programs. Based on historical experience, how big the balance should be in order to get through a recession is pretty predictable (in our case, about $1.2B). The only reason businesses don't howl as the trust fund balances go up in the case of UI is that the feds have stacked the deck: if a state doesn't operate a "conforming" UI program the businesses will be subject to a federal tax that is about three times higher than most states' current levy.
Add on top of all that their tax system, which relies almost heavily on income taxes (over half of their budget money comes from this). Every time the economy tanks, so does their revenue.
If you have a tax system that generates stable revenues despite the ups and downs of the economy, the stock market, and property prices, and is politically feasible, I know 50 states that would love to hear from you. Head taxes are stable but unaffordable for the poor (at least at the levels that would be required). Progressive consumption taxes are more stable than income tax, but still subject to revenue declines during a broad economic slump. Ditto for VATs, which also suffer from being regressive when you map it out to end consumers.
Personally, I'd prefer a progressive consumption tax. Essentially all of the information necessary to calculate it is already reported to the federal government for income tax purposes. But I don't think you could get it implemented: too many rich and powerful people would lose the favorable treatment they receive under the current tax codes.
Management is looking for you to bridge that gap between technology and business, which can either be bridged by you framing things into a business perspective or making them learn more about technology so they can do it. Your value to the company will increase if you take the former approach.
Second the first part of that.
Try to put yourself in their shoes:
If I were running this company, what would I need to know about the IT stuff?
"The sales staff is unable to access the order system for 15 minutes per week" is much more useful than uptime percentages.
The first is a business problem, the second is just geek-speak.
"The new desktop systems each save $150 per year in electricity costs" is more useful than the watts they use.
I recall one post-acquisition case where the big boss had to choose between
the two IT directors.
He kept the one from the acquired company and let go the one he had worked with for some years.
His description of it was "Bob never really came to grips with the idea that we were in business to sell widgets, not to employ an IT department."
The "Secure Grid 2009" security game run by DHS for senior policy makers identified the very high voltage transformers at the ends of those links as the weakest point.
These are so large (250 tons and up) that they can only be moved by rail, along lines with sufficient clearance.
For the most part, there are no spares and replacements require at least weeks and possibly months to build and transport.
Physically, most would be damaged sufficiently to fail if rammed with a garbage truck at reasonably high speed, and many are in positions where such an attack is possible.
A dozen such strikes, done at close to the same time, could conceivable take out much of the power supply for a state the size of Ohio for a period of several weeks.
Within several days, millions would be without water, sewage treatment, or reliable sources of food.
Not to defend Comcast particularly, but it is worth noting that as content aggregators, they are often caught in a bind not entirely of their own making. Ten years ago, (1) the FCC required that everything up through "enhanced basic" service be delivered without a cable box and (2) many content owners like ESPN required analog delivery of the signal as a part of their contracts. At that point in time, most content owners -- say, an NBC or CBS -- would not allow PVR-like delivery outside of the scheduled broadcast time. Much of that was driven by the content owners' business models, or by beliefs (true or not) held by the advertisers who were buying commercial slots.
At least at the cable company where I did research/engineering ten years ago, we were excited about all the new service opportunities that digital storage and switching made possible. The content owners, not nearly so much.
Until your bits reach the DSLAM or other aggregation device. Then your bits are multiplexed along with a bunch of other customers' bits, queued through whatever bandwidth AT&T has chosen to deploy. A hundred of you sharing a 50Mb/s downstream link serving a DSLAM deployed to your neighborhood have no advantage over a hundred cable modem customers sharing a 54Mb/s downstream link. And trust me on this one, both AT&T and Comcast are going to multiplex as many customers as they think they can get away with onto a shared link somewhere.
Within limits, expert systems seem to work reasonably well. Properly-trained software that examines x-ray images has been reported to have better accuracy than humans at diagnosing specific problems. The literature seems to suggest that expert systems for medical case diagnosis is more accurate than doctors and nurses, especially tired doctors and nurses. OTOH, patients have an intense dislike of such systems, particularly the diagnosis software, since it can seem like an arbitrary game of "20 Questions". Of course, these are tools that help the experts do their job better, not replacements for the expert people themselves.
Don't forget sharks, that seem to be fooled by the electric field that results from the DC current powering the repeaters, and occasionally attack the cables. I believe newer cables include upgraded armor that is more shark-resistant.
A Mac Mini draws about 85 Watts, so that isn't an option either.
Measuring the power draw of my (admittedly four years old) Mini, I have trouble getting it up to 65 watts. And when not doing anything in particular, it fairly quickly drops off into the 15-20 watt range. Averaged over the whole day, it comes in pretty close to your 30-watt target. But even the minimum configuration costs much more than you want to spend, new. I would also comment that I have been regularly frustrated when porting assorted software packages to OSX; I find myself doing more and more "UNIX" things in a VirtualBox VM running Ubuntu.
...hoping this will pave the way for "You want health insurance? We just need to sequence your genome first. Oh, sorry, you're going to get Huntingtons disease. Good luck with that."
At least for this group, doesn't matter. At age 65+, they're all eligible for the kind of can't-be-turned-down everyone-pays-the-same-premium government-operated socialized health insurance that Congress seems to think would be a disaster for the rest of us.
Also program guide information, changes in the subscriber's service, and in some cases, the boxes are polled individually about PPV or other premium services buys. Add in that the STB vendor is always under lots of pressure from the cable companies to shave a few bucks off the price of the box, and that the cable companies weren't paying for the power, and there's zero incentive to use low power (more expensive) parts or go to the trouble of having the controller shut off power to unneeded components.
Good quality voice communication has fairly low bandwidth requirements, but very tight latency limits. Above 20ms, you start to notice the lag. Above 50ms, it gets rather annoying. Beyond 150ms, you wouldn't want to use it for anything but absolute emergencies.
Indeed.
Many years ago,
when I did telephony testing work at Bell Labs,
the upper bound on round-trip delay was about 250 ms.
That's about the point where people begin "interpreting"
the delay in emotional terms.
One common form of that was business people's perception that
the other party was trying to be extremely careful
about their choice of words,
so was probably hiding something.
As a result,
when satellites were introduced,
only one direction of a two-way trunk
was carried on a satellite;
the other direction was terrestrial in order to keep round-trip delay down.
500 ms is the point where people decide that
the other person has not understood the question
and try to rephrase it.
It's always fun to watch an untrained news reporter
attempt to conduct an interview over
a two-way satellite connection,
where the latency is just over 500 ms.
IMO, his most interesting story looking at how the full spectrum of basics that a society has to provide -- air, food, shelter, marriage, child-rearing, allocation of scarce resources -- might change under suddenly different circumstances. The technology is comprehensible to almost anyone. And kids in high school today may live long enough to see computers that pass a Turing test -- certainly more likely to see that than FTL space flight. Start from "Does an AI that passes such a test have any rights?" and you can take the discussion anywhere.
...the problem is building quality tooling capable of machining the equipment required to refine the fissionable material...
While centrifuges suitable for uranium enrichment are precision devices,
a heavy water reactor capable of producing plutonium from un-enriched uranium
is much less so.
One seldom-mentioned part of Iran's program is construction of
a heavy-water separation plant and
a heavy-water reactor
estimated to be able to produce about two bombs worth of plutonium annually.
As to fabrication of an implosion-style fission device,
1940s machine tools were adequate, and
advances in readily-available automated precision machine tools in the last couple of decades
make it unlikely machining would be a limiting factor.
The Nth Country Experiment
in the mid-1960s was three people,
one of whom left fairly early,
all physics post-docs but none with weapons experience.
None were given access to classified information.
The conclusions were redacted when the original report was declassified,
but most experts seem to believe that
the group produced a workable design for an implosion-type device.
As a member of the NPT Iran is well within its rights to posses the outlined technologies.
As an NPT signatory, Iran is also obligated to open its entire nuclear program to inspection
in order to verify that no weapons development is happening.
The recent "Oh, did we forget to mention that 2nd centrifuge facility?" development
is almost certainly a violation.
At the very least,
it puts them in the uncomfortable position of having to
have the IAEA believe them when they say,
"Trust us, now we've shown you all the facilities."
You misunderstand my complaint. Investments in efficiency/conservation work today. Wind works today. Low-efficiency but cheap dye-sensitized solar panels work today. Kilowatt-class direct carbon fuel cells that can use biomass work in the lab today. Metal-air batteries with fast mechanical replacement and separate recharging appear to have sufficient energy and power densities for traction applications today. It's not that I'm unwilling to fund research that may bear fruit X years down the road, where X may or may not be within my lifetime; it's that we don't have X years.
It's just one person's opinion, but I believe that the limiting factor to deal with the energy problems will be capital; that is, the costs to replace or modify infrastructure and long-lived goods may be larger than we manage and sustain some version of the system. The choices of where we put that capital will be important, and research with uncertain benefits, or benefits that pay off only in the long term, may not be the right place.
But the scope on "yet" may also be important. Those on the "doomer" side of the peak energy debate would argue that "yet" needs to be within the next several years, or it won't matter, because we'll have started down a positive-feedback decline curve. If you knew with absolute certainty that, no matter how much research money you spent, cheap nanotubes were at least 70 years away, would you spend any money on that research, or would you spend it on something else?
I hate articles where they don't show their work. Assume that this figure was reached by dividing total electricity consumption in Japan by the number of households. Multiply the 3.4 kWh per hour by 24 and get 81.6 kWh per day. As a sanity check, compare that to the US: approximately 4e12 kWh consumed each year, 123e6 residential customers (both numbers from the EIA), 365 days in a year and... carry the one... 89.1 kWh per household per day. My suburban household consumes 25-26 kWh directly each day; but a more meaningful number also includes my share of the power at work, that consumed by the electric rail I ride to work on, the streetlights on my block, the refrigeration units at the grocery, the aluminum in my beer can, etc.
Yep. And in ten years or so, anyone with the budget of even a tiny country will be able to assemble designer viruses -- think Ebola with robust airborne transmission. If I'm going to worry, I worry about that a lot more than I worry about someone outside the existing nuclear club being able to fabricate a nuke.
I have a somewhat different perspective on the compiler issue. I run several pieces of older software in the underlying UNIX-like environment on my Mac. Written in Perl, using Perl/Tk for some GUI functions and Perl Data Language (PDL) with slatec for some of the number crunching. Snow Leopard comes with Perl 5.10.1 and some minimal supporting kit; that's okay, I don't expect the base distro for any UNIX-like system to include all of the Perl stuff. On every Ubuntu release I have tried, Perl/Tk has been available with apt-get, and PDL/slatec has built without difficulty. On Snow Leopard, Perl/Tk built but failed a lot of the tests. As Snow Leopard does not come with a functional Fortran, building PDL and the libraries it uses is not possible (and routine Makefile.PL files from CPAN don't find the Darwin port of gfortran that I used under Leopard).
For the time being, I'm using ActivePerl 5.8.9, for which Perl/Tk is available as a binary, and which runs under Snow Leopard. One of this fall's projects will be rewriting to use Tkx, which actually has multiple advantages. In the one case where I really need PDL, I'm running that code on Ubuntu in a virtual machine. And hoping that the Darwin port people eventually catch up.
It would be nice if Apple would either commit to providing a robust set of UNIX development tools, or none. This half-way-in-between Xcode arrangement is frustrating.
Slashdot Mirror
I taught myself Fortran 40 years ago when I was 16 (long story about how that happened that's not relevant here). I would have been so much better off if there had been someone looking over my shoulder from time to time and making suggestions about style. I estimate that it took two years of computer science in college before I unlearned all of the bad habits I picked up in three months...
I recall a question used (many years ago) on some OCS exams. A detailed list of available materials was supplied, along with a sketch of the terrain surrounding a portion of a stream. The question was "How would you build a bridge capable of carrying jeeps across this stream?" The correct answer was "Sergeant, take these men and this pile of stuff and build a bridge across this stream. I'll be back in three hours." Some incorrect answers did get people into various specialist training programs.
More than 50,000,000 annually. Automotive World reported that global commercial vehicle production for 2008 was 70.5 million units, including 52.6 million passenger cars, 13.6 million light commercial vehicles, and 3.6 million trucks. Total production was down 10.8% from 2007 due to the effects of the world-wide recession. Production in China and India for local sales are increasing very rapidly.
Indeed. For many people, the worst parts of previous generations' speculative fiction appears to be coming true.
I spent the last three years on the budget staff for a state legislature. I feel quite comfortable saying that it is almost impossible, politically, for states to accumulate a meaningful rainy-day fund in the contemporary political climate -- for a medium-sized state, say a billion or more. In my medium-sized state, a billion would have been about half as much as was/is needed to fill the "revenue gap" this recession. Once the surplus starts to pile up, there is enormous pressure to cut tax rates rather than grow the surplus to the appropriate size. We have an easy citizen initiative process; if the state accumulated a billion dollar reserve I can guarantee there would be a ballot initiative to refund much of that to taxpayers and the initiative would have a very good chance of passing. Polls show consistently that our voters believe they can have both high spending and low taxes and are frustrated that the elected officials don't deliver it.
An exception where states do manage to accumulate reasonable surpluses is in their unemployment insurance programs. Based on historical experience, how big the balance should be in order to get through a recession is pretty predictable (in our case, about $1.2B). The only reason businesses don't howl as the trust fund balances go up in the case of UI is that the feds have stacked the deck: if a state doesn't operate a "conforming" UI program the businesses will be subject to a federal tax that is about three times higher than most states' current levy.
If you have a tax system that generates stable revenues despite the ups and downs of the economy, the stock market, and property prices, and is politically feasible, I know 50 states that would love to hear from you. Head taxes are stable but unaffordable for the poor (at least at the levels that would be required). Progressive consumption taxes are more stable than income tax, but still subject to revenue declines during a broad economic slump. Ditto for VATs, which also suffer from being regressive when you map it out to end consumers.
Personally, I'd prefer a progressive consumption tax. Essentially all of the information necessary to calculate it is already reported to the federal government for income tax purposes. But I don't think you could get it implemented: too many rich and powerful people would lose the favorable treatment they receive under the current tax codes.
Second the first part of that. Try to put yourself in their shoes: If I were running this company, what would I need to know about the IT stuff? "The sales staff is unable to access the order system for 15 minutes per week" is much more useful than uptime percentages. The first is a business problem, the second is just geek-speak. "The new desktop systems each save $150 per year in electricity costs" is more useful than the watts they use. I recall one post-acquisition case where the big boss had to choose between the two IT directors. He kept the one from the acquired company and let go the one he had worked with for some years. His description of it was "Bob never really came to grips with the idea that we were in business to sell widgets, not to employ an IT department."
The "Secure Grid 2009" security game run by DHS for senior policy makers identified the very high voltage transformers at the ends of those links as the weakest point. These are so large (250 tons and up) that they can only be moved by rail, along lines with sufficient clearance. For the most part, there are no spares and replacements require at least weeks and possibly months to build and transport. Physically, most would be damaged sufficiently to fail if rammed with a garbage truck at reasonably high speed, and many are in positions where such an attack is possible. A dozen such strikes, done at close to the same time, could conceivable take out much of the power supply for a state the size of Ohio for a period of several weeks. Within several days, millions would be without water, sewage treatment, or reliable sources of food.
Not to defend Comcast particularly, but it is worth noting that as content aggregators, they are often caught in a bind not entirely of their own making. Ten years ago, (1) the FCC required that everything up through "enhanced basic" service be delivered without a cable box and (2) many content owners like ESPN required analog delivery of the signal as a part of their contracts. At that point in time, most content owners -- say, an NBC or CBS -- would not allow PVR-like delivery outside of the scheduled broadcast time. Much of that was driven by the content owners' business models, or by beliefs (true or not) held by the advertisers who were buying commercial slots.
At least at the cable company where I did research/engineering ten years ago, we were excited about all the new service opportunities that digital storage and switching made possible. The content owners, not nearly so much.
Until your bits reach the DSLAM or other aggregation device. Then your bits are multiplexed along with a bunch of other customers' bits, queued through whatever bandwidth AT&T has chosen to deploy. A hundred of you sharing a 50Mb/s downstream link serving a DSLAM deployed to your neighborhood have no advantage over a hundred cable modem customers sharing a 54Mb/s downstream link. And trust me on this one, both AT&T and Comcast are going to multiplex as many customers as they think they can get away with onto a shared link somewhere.
Within limits, expert systems seem to work reasonably well. Properly-trained software that examines x-ray images has been reported to have better accuracy than humans at diagnosing specific problems. The literature seems to suggest that expert systems for medical case diagnosis is more accurate than doctors and nurses, especially tired doctors and nurses. OTOH, patients have an intense dislike of such systems, particularly the diagnosis software, since it can seem like an arbitrary game of "20 Questions". Of course, these are tools that help the experts do their job better, not replacements for the expert people themselves.
Don't forget sharks, that seem to be fooled by the electric field that results from the DC current powering the repeaters, and occasionally attack the cables. I believe newer cables include upgraded armor that is more shark-resistant.
Measuring the power draw of my (admittedly four years old) Mini, I have trouble getting it up to 65 watts. And when not doing anything in particular, it fairly quickly drops off into the 15-20 watt range. Averaged over the whole day, it comes in pretty close to your 30-watt target. But even the minimum configuration costs much more than you want to spend, new. I would also comment that I have been regularly frustrated when porting assorted software packages to OSX; I find myself doing more and more "UNIX" things in a VirtualBox VM running Ubuntu.
At least for this group, doesn't matter. At age 65+, they're all eligible for the kind of can't-be-turned-down everyone-pays-the-same-premium government-operated socialized health insurance that Congress seems to think would be a disaster for the rest of us.
Also program guide information, changes in the subscriber's service, and in some cases, the boxes are polled individually about PPV or other premium services buys. Add in that the STB vendor is always under lots of pressure from the cable companies to shave a few bucks off the price of the box, and that the cable companies weren't paying for the power, and there's zero incentive to use low power (more expensive) parts or go to the trouble of having the controller shut off power to unneeded components.
Indeed. Many years ago, when I did telephony testing work at Bell Labs, the upper bound on round-trip delay was about 250 ms. That's about the point where people begin "interpreting" the delay in emotional terms. One common form of that was business people's perception that the other party was trying to be extremely careful about their choice of words, so was probably hiding something. As a result, when satellites were introduced, only one direction of a two-way trunk was carried on a satellite; the other direction was terrestrial in order to keep round-trip delay down.
500 ms is the point where people decide that the other person has not understood the question and try to rephrase it. It's always fun to watch an untrained news reporter attempt to conduct an interview over a two-way satellite connection, where the latency is just over 500 ms.
The Moon is a Harsh Mistress.
IMO, his most interesting story looking at how the full spectrum of basics that a society has to provide -- air, food, shelter, marriage, child-rearing, allocation of scarce resources -- might change under suddenly different circumstances. The technology is comprehensible to almost anyone. And kids in high school today may live long enough to see computers that pass a Turing test -- certainly more likely to see that than FTL space flight. Start from "Does an AI that passes such a test have any rights?" and you can take the discussion anywhere.
While centrifuges suitable for uranium enrichment are precision devices, a heavy water reactor capable of producing plutonium from un-enriched uranium is much less so. One seldom-mentioned part of Iran's program is construction of a heavy-water separation plant and a heavy-water reactor estimated to be able to produce about two bombs worth of plutonium annually. As to fabrication of an implosion-style fission device, 1940s machine tools were adequate, and advances in readily-available automated precision machine tools in the last couple of decades make it unlikely machining would be a limiting factor.
could someone furnish us with a link?
The Nth Country Experiment in the mid-1960s was three people, one of whom left fairly early, all physics post-docs but none with weapons experience. None were given access to classified information. The conclusions were redacted when the original report was declassified, but most experts seem to believe that the group produced a workable design for an implosion-type device.
As a member of the NPT Iran is well within its rights to posses the outlined technologies.
As an NPT signatory, Iran is also obligated to open its entire nuclear program to inspection in order to verify that no weapons development is happening. The recent "Oh, did we forget to mention that 2nd centrifuge facility?" development is almost certainly a violation. At the very least, it puts them in the uncomfortable position of having to have the IAEA believe them when they say, "Trust us, now we've shown you all the facilities."
You misunderstand my complaint. Investments in efficiency/conservation work today. Wind works today. Low-efficiency but cheap dye-sensitized solar panels work today. Kilowatt-class direct carbon fuel cells that can use biomass work in the lab today. Metal-air batteries with fast mechanical replacement and separate recharging appear to have sufficient energy and power densities for traction applications today. It's not that I'm unwilling to fund research that may bear fruit X years down the road, where X may or may not be within my lifetime; it's that we don't have X years.
It's just one person's opinion, but I believe that the limiting factor to deal with the energy problems will be capital; that is, the costs to replace or modify infrastructure and long-lived goods may be larger than we manage and sustain some version of the system. The choices of where we put that capital will be important, and research with uncertain benefits, or benefits that pay off only in the long term, may not be the right place.
But the scope on "yet" may also be important. Those on the "doomer" side of the peak energy debate would argue that "yet" needs to be within the next several years, or it won't matter, because we'll have started down a positive-feedback decline curve. If you knew with absolute certainty that, no matter how much research money you spent, cheap nanotubes were at least 70 years away, would you spend any money on that research, or would you spend it on something else?
I hate articles where they don't show their work. Assume that this figure was reached by dividing total electricity consumption in Japan by the number of households. Multiply the 3.4 kWh per hour by 24 and get 81.6 kWh per day. As a sanity check, compare that to the US: approximately 4e12 kWh consumed each year, 123e6 residential customers (both numbers from the EIA), 365 days in a year and... carry the one... 89.1 kWh per household per day. My suburban household consumes 25-26 kWh directly each day; but a more meaningful number also includes my share of the power at work, that consumed by the electric rail I ride to work on, the streetlights on my block, the refrigeration units at the grocery, the aluminum in my beer can, etc.
Yep. And in ten years or so, anyone with the budget of even a tiny country will be able to assemble designer viruses -- think Ebola with robust airborne transmission. If I'm going to worry, I worry about that a lot more than I worry about someone outside the existing nuclear club being able to fabricate a nuke.
I have a somewhat different perspective on the compiler issue. I run several pieces of older software in the underlying UNIX-like environment on my Mac. Written in Perl, using Perl/Tk for some GUI functions and Perl Data Language (PDL) with slatec for some of the number crunching. Snow Leopard comes with Perl 5.10.1 and some minimal supporting kit; that's okay, I don't expect the base distro for any UNIX-like system to include all of the Perl stuff. On every Ubuntu release I have tried, Perl/Tk has been available with apt-get, and PDL/slatec has built without difficulty. On Snow Leopard, Perl/Tk built but failed a lot of the tests. As Snow Leopard does not come with a functional Fortran, building PDL and the libraries it uses is not possible (and routine Makefile.PL files from CPAN don't find the Darwin port of gfortran that I used under Leopard).
For the time being, I'm using ActivePerl 5.8.9, for which Perl/Tk is available as a binary, and which runs under Snow Leopard. One of this fall's projects will be rewriting to use Tkx, which actually has multiple advantages. In the one case where I really need PDL, I'm running that code on Ubuntu in a virtual machine. And hoping that the Darwin port people eventually catch up.
It would be nice if Apple would either commit to providing a robust set of UNIX development tools, or none. This half-way-in-between Xcode arrangement is frustrating.