As I responded above, read the CAIR report or the Tufte papers on either the Challenger or Columbia accidents: in both cases, the engineers had identified the risks a priori (for Challenger, the decision to discount the effect on temperature on the O-rings was debated until the day before the launch). In both cases, the presentation was done in a very bad or CYA manner, and the result was that the risks were discounted and the managers made the wrong decisions.
Personally, I've been in a very similar situation, as a mere lieutenant: I was pressured (by a 3-star general) to not issue a grounding order for a hundred or so aircraft. I simply said that as an engineer I had to make a correct technical recommendation and the final decision was up to management. In the end, the order stood.
Read the CAIR report or the Tufte paper on it (the famous "Powerpoint is harmful" paper); the engineers did try to show to their managers how important the foam impact was; but they covered their behinds too much, made a bad presentation, and as a result the not-too-technical managers discounted the importance of the impact. The bureaucracy in this case worked; the engineers failed.
As for your comment on culture, I agree with your thought but disagree with the conclusion: the whole point of a bureaucracy such as NASA's is to minimize risk, not maximize profit/reward. In engineering the risks/innovation should be done at the design stage, not during implementation, maintainance or operations.
As a former Army Aviation maintainance bureaucrat, I can safely say that in your example above the fault lies with the engineer(s) that submitted the second request: if you are authorized for such a request and are denied you can still appeal through proper bureaucratic channels up the chain of command.
If that fails, then clearly the error lies with the person that made the second or N+1 request: the Air Force is not in the business of losing spacecraft or astronauts: if the importance of getting those pictures was clearly shown, there is no way that any reasonable officer would have denied it.
Bottom line: bureaucracies don't fail, people do (because they can always work the system). There is no such thing as a no-fault error in engineering.
OK, this is a plug, but I think memigo does a mobile news portal genuinely better: you get a personalized URL to access news over a mobile phone or PDA, and what's more the headline view can link you to only lightly-formatted content of each article (i.e. a PDA- or printer-formatted version of each article). Much better than linking to whatever version the site's RSS feed links to. Check it out.
Yes, also the author left out the -O optimization, claiming that that it only optimizes for loading. That's absolutely true, so I hope he included compilation times in the benchmarks for C/C++/Java (he didn't).
Comparing execution of a compiled, low-level language to syntax-checking/optimization/writing of bytecode to disk/execution of a dynamic languages is insane.
It's been a while since I played with Beowulf clusters, but I've been thinking of using a Beo-class cluster in an enterprise app. The problem there is that you want the IT people to have to do as little "deep magic" work as possible when it comes to grids. So, auto-detection of new nodes, transfer of jobs out of failing/crashed nodes, etc. are important features. Is there anything close to that for Linux? a few different solutions that integrate well together isn't out of the question either...
Search engines index words in web pages and at most keep word order (word no X out of N in page). In addition, they throw out common or garbage words (like 'the', 'and', etc), so the actual text is not represented in the index. That right there is a major reason why not to do regexes... Secondly, regexes are awefully expensive CPU-wise and understood by only a tiny portion of your users, so it's quite unlikely that this would happen.
Well, for an attempt at a better newsbot than Google news, you can check out newsbot here. It does a few things that GN leaves out (XML feeds, PDA version, peer recommendations, etc, etc) and I believe it has a better S/N ratio. End of shameless plug.
I travelled 150k+ miles a year across the globe until recently. Here's a quick how-to on travelling with gadgets (laptop, multiple adapters/converters, a PDA and at least one cellphone):
Stuff almost everything in a small carry-on bag. A professional-looking backpack is the best (the backpack part for ergonomic, not security reasons).
Wear as little metal as possible. That includes shoes with metal inserts in the soles (a lot of plain-looking walking or dress shoes have those). Get a coin purse and stuff your change in the purse and that in your carry-on.
When going thru security, do not rush, follow instructions even inane ones ("yes, I will turn on that Palm V for you sir"). Do not tell them that you got a flight to catch or attract their attention in any other way. Since you're not carrying anything dangerous (right?) it will be far faster for you to go with the flaw and accept the default process rather than try to explain your reasons for short-circuiting it.
They stopped requiring booting laptops a while back. Most US airports I've flown thru use instead chemical "sniffer" devices --they wipe your laptop with a sticky pad and then scan the pad in the sniffer machine. Much better than a simple boot...
Agreed; I've travelled a lot internationally and my favorite solution to this is in place at Hong Kong's airport: security gives you a big laminated piece of paper for each tray you put thru the X-rays, with a number that matches another piece of paper laminated to the bottom of said tray. On the other side of the machine, you can only pick up the tray(s) that you have the token(s) for. Simple, quick and pretty secure.
The real problem isn't the polling intervals, is that most RSS readers/spiders do not respect HTTP 304 (Not Modified). RSS is ideal for Etag/Not-Modified-Since behavior, but no, most spiders are still too lazy to implement this.
My newsbot (in my.sig) creates dynamic RSS feeds, customized for each agent; I thought that was a great feature to give users, but it's getting overused by some spiders hitting the site every 15-20 minutes, w/o listening for 304s...
Shameless plug ahead: my home-brew newsbot predates both MSN Newsbot and Google News and has way more features (incl. collaborative filtering, news alerts, customized XML and Palm feeds and more).
I am surprised that no one yet has pointed out the ramifications of Jobs' revelation: a loss leader is an anti-competitive practice --it's actually illegal in a few countries, although nowhere in the US AFAIK. The fact that Apple can afford to sell songs for 99c/song because they make up their operating profits on iPod sales means two things:
1. No company that does not sell hardware can afford to compete with those that do in the online music store business (i.e. Apple, Dell, Samsung kinda).
2. The prices of the aforementioned units that subsidize the stores are artificially inflated by the amortized costs of the online music store operations, or expected growth. I.e. the iPods are indeed too expensive.
Either way, this does suck. And it sucks because the magic price-point.
Which part of the audio CD do you like? that you can misplace it, toast in a microwave or bake it in your car? or that you're likely to never examing the leaflet/cover art?
I agree with you to a point, but the same arguments hold true for newspapers or books. Digital technology will eventually replace all of this media, despite our emotional attachments to more tangible representations, for two reasons: convenience and price. The convenience factor is already here for news and music (books are still a bit far off) and as for price, well, when the business models get ironed out (it seems like iTunes is on its way at least), the prices will drop to more attractive levels.
But don't mourn the passing of less efficient media. I for one would rather have say a Flash animation instead of a CD booklet any day. At least I wouldn't have to hunt for it when I wanna lookup a lyric...
[Plug] My newsbot does News Search alerts plus context alerts (news articles it thinks you may like based on your past read/rated articles) and they are all available via customized RSS and PDA feeds.
Python is a refactorer's dream. You can transition your Java application to Jython re-using your Java classes while ironing out the bugs and design of the Python code, implementing caching, static HTML generation and the like.
When you're done, swap the JVM out of Jython and run pure Python with debugged code. If Python gives you any performance trouble, write small C-based modules for your frequently used code and wrap it in Python (fairly easy to do).
Loops, if blocks and the like are marked by identation/whitespace. Put another way, there are no {} religious wars, coz there ain't any {}s. Trust me, it's a beautiful thing.
Code makes sense the first time you look at it. There are a few Python idioms that will take a bit if you're coming from the outside, but you will learn them fast and they are not that complicated.
You will never appreciate another language again. Proceed with caution, this may be detrimental to your employment:-)
Sample code:
myfile = open("data.csv") for line in myfile.readlines():
columns = line.split(',')
print "Column 2 is %s " % columns[1]
myfile.close()
Can you tell what this does? This is not the exception, this is the rule. Even deep-magic Python code is understandable by another developer after just a few looks. That's the main, big advantage over other high-level languages...
It's not a panacea, agreed. However, the ability to track every individual product through the supply chain will indeed revolutionize supply chain management. It will make every major retailer as logistically aware as UPS is now of every package in their system. Which, given good enough smart data-mining and decision-support software will indeed cut down costs a lot and end up lowering prices for the end-consumer. The trouble of course is a) dealing with additional TBs of data *per day*, b) absorbiing the massive infrastructure costs to even come close using the TBs of data.
Of course, I am biased, I am a retail DSS consultant:-)
My guess is that the abundance of data from EPC/RFID/whatever will make supply-chain-sensitive retailers (like grocers or discounters) that much more competitive and will force even greater consolidation. Fashion retailers and the like are probably not going to care as much (success in fashion retail is much less dependent on supply chain mgt).
Slashdot Mirror
As I responded above, read the CAIR report or the Tufte papers on either the Challenger or Columbia accidents: in both cases, the engineers had identified the risks a priori (for Challenger, the decision to discount the effect on temperature on the O-rings was debated until the day before the launch). In both cases, the presentation was done in a very bad or CYA manner, and the result was that the risks were discounted and the managers made the wrong decisions.
Personally, I've been in a very similar situation, as a mere lieutenant: I was pressured (by a 3-star general) to not issue a grounding order for a hundred or so aircraft. I simply said that as an engineer I had to make a correct technical recommendation and the final decision was up to management. In the end, the order stood.
Read the CAIR report or the Tufte paper on it (the famous "Powerpoint is harmful" paper); the engineers did try to show to their managers how important the foam impact was; but they covered their behinds too much, made a bad presentation, and as a result the not-too-technical managers discounted the importance of the impact. The bureaucracy in this case worked; the engineers failed.
As for your comment on culture, I agree with your thought but disagree with the conclusion: the whole point of a bureaucracy such as NASA's is to minimize risk, not maximize profit/reward. In engineering the risks/innovation should be done at the design stage, not during implementation, maintainance or operations.
As a former Army Aviation maintainance bureaucrat, I can safely say that in your example above the fault lies with the engineer(s) that submitted the second request: if you are authorized for such a request and are denied you can still appeal through proper bureaucratic channels up the chain of command.
If that fails, then clearly the error lies with the person that made the second or N+1 request: the Air Force is not in the business of losing spacecraft or astronauts: if the importance of getting those pictures was clearly shown, there is no way that any reasonable officer would have denied it.
Bottom line: bureaucracies don't fail, people do (because they can always work the system). There is no such thing as a no-fault error in engineering.
IBM and Philips are also getting into the action.
OK, this is a plug, but I think memigo does a mobile news portal genuinely better: you get a personalized URL to access news over a mobile phone or PDA, and what's more the headline view can link you to only lightly-formatted content of each article (i.e. a PDA- or printer-formatted version of each article). Much better than linking to whatever version the site's RSS feed links to. Check it out.
MP3s? The first true music-swapping service was sounds.sdsu.edu, wayyyyy back in '92 or so...
Yes, also the author left out the -O optimization, claiming that that it only optimizes for loading. That's absolutely true, so I hope he included compilation times in the benchmarks for C/C++/Java (he didn't).
Comparing execution of a compiled, low-level language to syntax-checking/optimization/writing of bytecode to disk/execution of a dynamic languages is insane.
It's been a while since I played with Beowulf clusters, but I've been thinking of using a Beo-class cluster in an enterprise app. The problem there is that you want the IT people to have to do as little "deep magic" work as possible when it comes to grids. So, auto-detection of new nodes, transfer of jobs out of failing/crashed nodes, etc. are important features. Is there anything close to that for Linux? a few different solutions that integrate well together isn't out of the question either...
Search engines index words in web pages and at most keep word order (word no X out of N in page). In addition, they throw out common or garbage words (like 'the', 'and', etc), so the actual text is not represented in the index. That right there is a major reason why not to do regexes... Secondly, regexes are awefully expensive CPU-wise and understood by only a tiny portion of your users, so it's quite unlikely that this would happen.
Well, for an attempt at a better newsbot than Google news, you can check out newsbot here. It does a few things that GN leaves out (XML feeds, PDA version, peer recommendations, etc, etc) and I believe it has a better S/N ratio. End of shameless plug.
Stuff almost everything in a small carry-on bag. A professional-looking backpack is the best (the backpack part for ergonomic, not security reasons).
Wear as little metal as possible. That includes shoes with metal inserts in the soles (a lot of plain-looking walking or dress shoes have those). Get a coin purse and stuff your change in the purse and that in your carry-on.
When going thru security, do not rush, follow instructions even inane ones ("yes, I will turn on that Palm V for you sir"). Do not tell them that you got a flight to catch or attract their attention in any other way. Since you're not carrying anything dangerous (right?) it will be far faster for you to go with the flaw and accept the default process rather than try to explain your reasons for short-circuiting it.
They stopped requiring booting laptops a while back. Most US airports I've flown thru use instead chemical "sniffer" devices --they wipe your laptop with a sticky pad and then scan the pad in the sniffer machine. Much better than a simple boot...
Agreed; I've travelled a lot internationally and my favorite solution to this is in place at Hong Kong's airport: security gives you a big laminated piece of paper for each tray you put thru the X-rays, with a number that matches another piece of paper laminated to the bottom of said tray. On the other side of the machine, you can only pick up the tray(s) that you have the token(s) for. Simple, quick and pretty secure.
Are there any other such devices that pick up internet radio (Shoutcast-only is fine) besides MP3 playing? I don't think the Slimp3 does radio, right?
Well, my newsbot does a lot of what you describe, at least for news articles. Give it a shot.
This is my newsbot.
The real problem isn't the polling intervals, is that most RSS readers/spiders do not respect HTTP 304 (Not Modified). RSS is ideal for Etag/Not-Modified-Since behavior, but no, most spiders are still too lazy to implement this.
.sig) creates dynamic RSS feeds, customized for each agent; I thought that was a great feature to give users, but it's getting overused by some spiders hitting the site every 15-20 minutes, w/o listening for 304s...
My newsbot (in my
Shameless plug ahead: my home-brew newsbot predates both MSN Newsbot and Google News and has way more features (incl. collaborative filtering, news alerts, customized XML and Palm feeds and more).
Hit submit by mistake: ... and it sucks because the magic price point of 99c is made too expensive by the RIAA.
I am surprised that no one yet has pointed out the ramifications of Jobs' revelation: a loss leader is an anti-competitive practice --it's actually illegal in a few countries, although nowhere in the US AFAIK. The fact that Apple can afford to sell songs for 99c/song because they make up their operating profits on iPod sales means two things:
1. No company that does not sell hardware can afford to compete with those that do in the online music store business (i.e. Apple, Dell, Samsung kinda).
2. The prices of the aforementioned units that subsidize the stores are artificially inflated by the amortized costs of the online music store operations, or expected growth. I.e. the iPods are indeed too expensive.
Either way, this does suck. And it sucks because the magic price-point.
Which part of the audio CD do you like? that you can misplace it, toast in a microwave or bake it in your car? or that you're likely to never examing the leaflet/cover art?
I agree with you to a point, but the same arguments hold true for newspapers or books. Digital technology will eventually replace all of this media, despite our emotional attachments to more tangible representations, for two reasons: convenience and price. The convenience factor is already here for news and music (books are still a bit far off) and as for price, well, when the business models get ironed out (it seems like iTunes is on its way at least), the prices will drop to more attractive levels.
But don't mourn the passing of less efficient media. I for one would rather have say a Flash animation instead of a CD booklet any day. At least I wouldn't have to hunt for it when I wanna lookup a lyric...
[Plug] My newsbot does News Search alerts plus context alerts (news articles it thinks you may like based on your past read/rated articles) and they are all available via customized RSS and PDA feeds.
Python is a refactorer's dream. You can transition your Java application to Jython re-using your Java classes while ironing out the bugs and design of the Python code, implementing caching, static HTML generation and the like.
When you're done, swap the JVM out of Jython and run pure Python with debugged code. If Python gives you any performance trouble, write small C-based modules for your frequently used code and wrap it in Python (fairly easy to do).
Sample code:
myfile = open("data.csv")
for line in myfile.readlines():
columns = line.split(',')
print "Column 2 is %s " % columns[1]
myfile.close()
Can you tell what this does? This is not the exception, this is the rule. Even deep-magic Python code is understandable by another developer after just a few looks. That's the main, big advantage over other high-level languages...
It's not a panacea, agreed. However, the ability to track every individual product through the supply chain will indeed revolutionize supply chain management. It will make every major retailer as logistically aware as UPS is now of every package in their system. Which, given good enough smart data-mining and decision-support software will indeed cut down costs a lot and end up lowering prices for the end-consumer. The trouble of course is a) dealing with additional TBs of data *per day*, b) absorbiing the massive infrastructure costs to even come close using the TBs of data.
:-)
Of course, I am biased, I am a retail DSS consultant
My guess is that the abundance of data from EPC/RFID/whatever will make supply-chain-sensitive retailers (like grocers or discounters) that much more competitive and will force even greater consolidation. Fashion retailers and the like are probably not going to care as much (success in fashion retail is much less dependent on supply chain mgt).