I don't mean to be pedantic, but Python is inherently strongly typed. What it is not is "statically typed". I think there is a general confusion in this thread between strong vs weak typing, and dynamic vs static typing.
How do you handle complicated math that require special packages in DocBook?
What do you write Docbook in? Surely not direct XML?
I find myself using LaTeX for print documents, and Sphinx for multiformat documents. But I'd be interested to find out if a Docbook based workflow is actually practical for writing mathematical documents.
Actually that does not include Hong Kong. The report says Hong Kong is a moderate proficiency state, while India and China are both low proficiency nations. However the original point I was trying to make was that on average, China and India aren't too far apart in terms of English proficiency, despite what most people believe.
Of course, the report is talking about averages (I suspect the standard deviations are quite large). In terms of the number of *proficient* speakers in English in those countries, my sense is that India > Hong Kong > China.
p.s. in my experience, most Hong Kong people actually speak English poorly or not at all. The folks you met may have been from certain echelons of society that happened to have had good English instruction. The only nation with a majority Chinese population that speaks English reasonably well is Singapore, and even there, the distribution of fluency is largely skewed toward the highly educated.
Conventional wisdom has it that China lags India in English proficiency, for obvious reaons. However, this report says:
"Asia’s English proficiency scores show that reputations are not always accurate. Take for example the nearly equivalent scores of China and India. Despite its British colonial legacy and reputation as an English-speaking nation, India is today no more proficient in English than rapidly improving China."
Are you sure you want to do that? I can understand typesetting math in the browser, but typesetting entire TeX documents? There's already an AMS-endorsed way of typesetting TeX math (Javascript-based) called MathJax (http://www.mathjax.org/), and it works pretty well (well enough for sites like http://mathoverflow.net./
What puzzles me is that there is no confirmation step required in these contactless payment systems. When I buy stuff with my chip-based debit or credit card, I'm asked to enter a PIN. Else, I have to physically swipe the card to ensure there is no ambiguity as to whether or not I meant to pay with my card of choice.
With a contactless system, I could be wanting to pay with my credit card, but if I accidentally held my cell phone too close to the reader, it would debit the amount from my phone instead of my card. Why can't there be a screen that pops-up on the phone that says "Touch button to confirm payment"? This seems to me to be a major design flaw.
It's frustrating for us though when you air your documentaries in Canada, and are quoting ounces, Fahrenheit, yards, etc, since I honestly have no clue what you are talking about. I think it would be a nice gesture for us if you could at least subtitle the imperial measurements in metric or use both, if you must.
Actually Canada isn't as metric as you think. Due to our proximity to the U.S. (and our historical use of Imperial units), we've adopted a kind of a schizophrenic approach to units, and we've grown comfortable with it. Yes, we measure temperature in Celsius instead of Fahrenheit, but I'm sure you've noticed that produce/meat/fish are quoted in lbs (some grocery stores use kg, and a lot of people are thrown off by that). We measure distances in meters and kilometers, but colloquially, we say a person is 6"2' 180lbs (very few people know their height and weight in cm and kg). Our air conditioners are rated in Btu's rather than Watts. Canadian football fields are measured in yards. We buy 2 x 4s from Home Depot. And while our store bought beverages are in 350ml packages, at a bar we buy our beer in pints. Flat screen TVs? The Best Buy brochure says they're 52" instead of 132.08 cm.
In engineering, imperial units are still widely used. In engineering school, we spent 1/3 of a course in first year becoming familiar with both the SI and Imperial systems, and learning to convert between them (i.e. dimensional analysis... it's not as trivial as you think when you have to convert vapour and liquid compound properties, e.g. from SCFM to m3/s, you have to know what the standard conditions are). I think personally it's great that Canadian engineering graduates are trained in both systems.
The fact is, imperial units are just more natural for some things and less so for others. The same can be said of metric... especially for very small or very large quantities (e.g. Intel's 45nm process instead of 1.77165354e-6" process).
Most snow melters work at very high thermal efficiencies (90 - 98%). Typically, one ton of snow requires 1.5 US gallon of diesel to melt. Remember, snow is not ice -- it's far less dense. http://www.snowmelter.com/en/snowmelters_faq.php
Snow melters can melt anything from 20 to 5000 tons of snow/hour, depending on their design capacity. Airports already use this technology extensively -- it's nothing new. http://www.snowmelter.com/en/clients.php
When I read the phrase English Shell Code, the first thought that came to my mind was:
% Oh I say, can I see a list of files, old top?
-rwxrwxrwx 1 alfred staff 192 7 Mar 2008 teacosy.txt drwxr-xr-x 37 alfred staff 1258 25 Nov 2008 cricketscores -rwxr-xr-x 1 alfred staff 260 28 Aug 2008 cucumbersandwiches.py
% Spiffing, just spiffing. Shall we have a peek at the processes?
I did read the summary. I passed on information on a remote wipe service, which is one of the many options for doing what the poster wanted. What part of the summary did you have trouble understanding?
Well, all that would be unnecessary if server-side gzip were turned on. I consider that a type of web page optimization, and you don't really have to anything special with the HTML.
I believe there is a case to be made for compression even for very dynamic websites. It works very well for mobile devices like Blackberries.
I use Mercurial on Linux and OS X. I've found that for medium-sized projects, it is comparable to Subversion in terms of speed. But it is miles ahead in terms of ease of use.
I realize ease of use is a subjective thing -- svn would probably be intuitive to someone who's been raised on cvs. But for folks like me who have never used cvs before, Mercurial's design feels much cleaner.
The name "University of California", when unqualified, refers to the Berkeley campus.
It's just a convention. Other examples: 1) University of Michigan -> Ann Arbor campus 2) University of Wisconsin -> Madison campus 3) University of Illinois -> Urbana-Champaign campus 4) University of Maryland -> College Park campus
I agree, and I'm not sure boot times are all that relevant these days.
I usually just put my notebook to sleep instead of shutting it down. (the only times I shut it down are for system updates and stuff). So, it's "instant-on" for me.... but then again I have an Apple notebook. But I suspect it's the same with most other notebooks.
Even on my Dell desktop (running Ubuntu 8.04), it only takes 3-4 seconds to wake up from sleep.
I'm not sure that's true of Photoshop's target demographic.
Many people I know who use Photoshop (i.e. people who actually pay for licenses) often also use other pre-press software that aren't available on Linux. One would have to port the other tools too, and deal with lack of availability of drivers for special equipment. Photoshop is only one tool in the pre-press production chain. Hence the inertia.
I'm a Debian user myself, but I personally agree with the GP that the target market is just too small.
I take your point, but I just have a few to add: (I'm a PhD student too) 1) In MATLAB, you can actually look at the source and figure out what it's doing (unless it's a MEX routine). To use your example, you can actually type "edit ode45" in the console and the source code pops up. 2) It is not that difficult to use a MATLAB routine in a C or FORTRAN program. You just have to do a MEX callback. It's not entirely straightforward, but it's not exactly crazy either. There are times when you do want to do this: MATLAB has very robust and optimized routines for stuff like SVD and QR decomposition -- to code those up yourself would be a waste of time and you'd never get something as good what MATLAB has anyway. (hmmm.... though SVD and QR are bad examples -- those routines are actually available in LAPACK. But you get my drift.) 3) The quality of the algorithms in the much touted Numerical Recipes is haphazard. If you want the best algorithms (that handles degenerate cases etc.), you have to piece things together from various papers and do a lot of testing and add heuristics etc., which involves a lot of work, and you'd probably do a bad job of it anyway unless it's your area of specialty. For instance, anyone can implement a mixed integer linear solver (it's just a branch and bound algorithm). But the beauty of commercial software is that experts working on it have designed all kinds of special heuristics and speed-up methods that a dillentante could never think of -- which is why CPLEX (a commercial solver) works better than any of the open source options out there. 4) As for not knowing how your results were produced, it's a tradeoff we have to live with. We should aim for transparency in packages, but ultimately I think one should keep the final goal in mind: to produce good research. For me, that means using the right tools for the job, whether proprietary or free (and in certain fields, most of the good tools are proprietary). It means not reinventing the wheel. 5) Vendor lock-in is always a problem, and one of the ways I've overcome it is to write a DSL to represent my problems. From this DSL, I generate code in whatever language I need to work in. This takes a bit of upfront work, but if you do it right, it can save you a lot of time in the long run. I've found that this route has worked pretty well for me.... YMMV.
Slashdot Mirror
The Automat concept has been around for a long time
https://en.wikipedia.org/wiki/...
I don't mean to be pedantic, but Python is inherently strongly typed.
What it is not is "statically typed".
I think there is a general confusion in this thread between strong vs weak typing, and dynamic vs static typing.
I think you are complaining about voice recognition, not translation. Translation is mapping one language to another.
Names...
This reminds me of Spanish-French AI character in Space Quest 6 that would pop up when Roger Wilco pressed the autopilot button.
His name was "Manuel Auxveride". :)
How do you handle complicated math that require special packages in DocBook?
What do you write Docbook in? Surely not direct XML?
I find myself using LaTeX for print documents, and Sphinx for multiformat documents. But I'd be interested to find out if a Docbook based workflow is actually practical for writing mathematical documents.
I find I'm very productive when I focus on short tasks and switch between them (sort of like how co-routines work).
I'm not productive when I'm doing more than one thing at a time.
No, it's spelled globalization in Canada. Canadian English is a hybrid of British and American English. We do spell color as colour though....
Actually that does not include Hong Kong. The report says Hong Kong is a moderate proficiency state, while India and China are both low proficiency nations. However the original point I was trying to make was that on average, China and India aren't too far apart in terms of English proficiency, despite what most people believe.
Of course, the report is talking about averages (I suspect the standard deviations are quite large). In terms of the number of *proficient* speakers in English in those countries, my sense is that India > Hong Kong > China.
p.s. in my experience, most Hong Kong people actually speak English poorly or not at all. The folks you met may have been from certain echelons of society that happened to have had good English instruction. The only nation with a majority Chinese population that speaks English reasonably well is Singapore, and even there, the distribution of fluency is largely skewed toward the highly educated.
Conventional wisdom has it that China lags India in English proficiency, for obvious reaons. However, this report says:
"Asia’s English proficiency scores show that reputations are not always accurate. Take for example the nearly equivalent scores of China and India. Despite its British colonial legacy and reputation as an English-speaking nation, India is today no more proficient in English than rapidly improving China."
http://www.ef.com/sitecore/__/~/media/efcom/epi/pdf/EF-EPI-2011.pdf?ctr=ca
Are you sure you want to do that? I can understand typesetting math in the browser, but typesetting entire TeX documents?
There's already an AMS-endorsed way of typesetting TeX math (Javascript-based) called MathJax (http://www.mathjax.org/), and it works pretty well (well enough for sites like http://mathoverflow.net./
What puzzles me is that there is no confirmation step required in these contactless payment systems.
When I buy stuff with my chip-based debit or credit card, I'm asked to enter a PIN. Else, I have to physically swipe the card to ensure there is no ambiguity as to whether or not I meant to pay with my card of choice.
With a contactless system, I could be wanting to pay with my credit card, but if I accidentally held my cell phone too close to the reader, it would debit the amount from my phone instead of my card. Why can't there be a screen that pops-up on the phone that says "Touch button to confirm payment"? This seems to me to be a major design flaw.
It's frustrating for us though when you air your documentaries in Canada, and are quoting ounces, Fahrenheit, yards, etc, since I honestly have no clue what you are talking about. I think it would be a nice gesture for us if you could at least subtitle the imperial measurements in metric or use both, if you must.
Actually Canada isn't as metric as you think. Due to our proximity to the U.S. (and our historical use of Imperial units), we've adopted a kind of a schizophrenic approach to units, and we've grown comfortable with it. Yes, we measure temperature in Celsius instead of Fahrenheit, but I'm sure you've noticed that produce/meat/fish are quoted in lbs (some grocery stores use kg, and a lot of people are thrown off by that). We measure distances in meters and kilometers, but colloquially, we say a person is 6"2' 180lbs (very few people know their height and weight in cm and kg). Our air conditioners are rated in Btu's rather than Watts. Canadian football fields are measured in yards. We buy 2 x 4s from Home Depot. And while our store bought beverages are in 350ml packages, at a bar we buy our beer in pints. Flat screen TVs? The Best Buy brochure says they're 52" instead of 132.08 cm.
In engineering, imperial units are still widely used. In engineering school, we spent 1/3 of a course in first year becoming familiar with both the SI and Imperial systems, and learning to convert between them (i.e. dimensional analysis... it's not as trivial as you think when you have to convert vapour and liquid compound properties, e.g. from SCFM to m3/s, you have to know what the standard conditions are). I think personally it's great that Canadian engineering graduates are trained in both systems.
The fact is, imperial units are just more natural for some things and less so for others. The same can be said of metric... especially for very small or very large quantities (e.g. Intel's 45nm process instead of 1.77165354e-6" process).
Oops I meant 500 tons/hr, not 5000.
Most snow melters work at very high thermal efficiencies (90 - 98%). Typically, one ton of snow requires 1.5 US gallon of diesel to melt. Remember, snow is not ice -- it's far less dense.
http://www.snowmelter.com/en/snowmelters_faq.php
Snow melters can melt anything from 20 to 5000 tons of snow/hour, depending on their design capacity. Airports already use this technology extensively -- it's nothing new.
http://www.snowmelter.com/en/clients.php
When I read the phrase English Shell Code, the first thought that came to my mind was:
% Oh I say, can I see a list of files, old top?
-rwxrwxrwx 1 alfred staff 192 7 Mar 2008 teacosy.txt
drwxr-xr-x 37 alfred staff 1258 25 Nov 2008 cricketscores
-rwxr-xr-x 1 alfred staff 260 28 Aug 2008 cucumbersandwiches.py
% Spiffing, just spiffing. Shall we have a peek at the processes?
PID TTY TIME CMD
380 ttys000 0:00.01 -bash
I did read the summary. I passed on information on a remote wipe service, which is one of the many options for doing what the poster wanted.
What part of the summary did you have trouble understanding?
Website: http://www.absolute.com/products/lojack
FAQ: http://www.absolute.com/resources/public/FAQ/L4L-FAQ-E.pdf
Costs $59.95/year for the premium package which supports Remote Wipe. Embeds itself in the BIOS/EFI. Supports XP and OS X.
Well, all that would be unnecessary if server-side gzip were turned on. I consider that a type of web page optimization, and you don't really have to anything special with the HTML.
I believe there is a case to be made for compression even for very dynamic websites. It works very well for mobile devices like Blackberries.
I use Mercurial on Linux and OS X. I've found that for medium-sized projects, it is comparable to Subversion in terms of speed. But it is miles ahead in terms of ease of use.
I realize ease of use is a subjective thing -- svn would probably be intuitive to someone who's been raised on cvs. But for folks like me who have never used cvs before, Mercurial's design feels much cleaner.
The name "University of California", when unqualified, refers to the Berkeley campus.
It's just a convention. Other examples:
1) University of Michigan -> Ann Arbor campus
2) University of Wisconsin -> Madison campus
3) University of Illinois -> Urbana-Champaign campus
4) University of Maryland -> College Park campus
I agree, and I'm not sure boot times are all that relevant these days.
I usually just put my notebook to sleep instead of shutting it down. (the only times I shut it down are for system updates and stuff). So, it's "instant-on" for me.... but then again I have an Apple notebook. But I suspect it's the same with most other notebooks.
Even on my Dell desktop (running Ubuntu 8.04), it only takes 3-4 seconds to wake up from sleep.
Similar results from Hamilton, Ontario, Canada.
17 hops -> 12 hops
90 ms -> 25 ms
I'm not sure that's true of Photoshop's target demographic.
Many people I know who use Photoshop (i.e. people who actually pay for licenses) often also use other pre-press software that aren't available on Linux. One would have to port the other tools too, and deal with lack of availability of drivers for special equipment. Photoshop is only one tool in the pre-press production chain. Hence the inertia.
I'm a Debian user myself, but I personally agree with the GP that the target market is just too small.
I think the abstract bit refers to the art of the surrealist artist Joan Miró.
http://en.wikipedia.org/wiki/Joan_Miró
I'm surprised nobody mentioned him.
I take your point, but I just have a few to add: (I'm a PhD student too)
1) In MATLAB, you can actually look at the source and figure out what it's doing (unless it's a MEX routine). To use your example, you can actually type "edit ode45" in the console and the source code pops up.
2) It is not that difficult to use a MATLAB routine in a C or FORTRAN program. You just have to do a MEX callback. It's not entirely straightforward, but it's not exactly crazy either. There are times when you do want to do this: MATLAB has very robust and optimized routines for stuff like SVD and QR decomposition -- to code those up yourself would be a waste of time and you'd never get something as good what MATLAB has anyway. (hmmm.... though SVD and QR are bad examples -- those routines are actually available in LAPACK. But you get my drift.)
3) The quality of the algorithms in the much touted Numerical Recipes is haphazard. If you want the best algorithms (that handles degenerate cases etc.), you have to piece things together from various papers and do a lot of testing and add heuristics etc., which involves a lot of work, and you'd probably do a bad job of it anyway unless it's your area of specialty.
For instance, anyone can implement a mixed integer linear solver (it's just a branch and bound algorithm). But the beauty of commercial software is that experts working on it have designed all kinds of special heuristics and speed-up methods that a dillentante could never think of -- which is why CPLEX (a commercial solver) works better than any of the open source options out there.
4) As for not knowing how your results were produced, it's a tradeoff we have to live with. We should aim for transparency in packages, but ultimately I think one should keep the final goal in mind: to produce good research. For me, that means using the right tools for the job, whether proprietary or free (and in certain fields, most of the good tools are proprietary). It means not reinventing the wheel.
5) Vendor lock-in is always a problem, and one of the ways I've overcome it is to write a DSL to represent my problems. From this DSL, I generate code in whatever language I need to work in. This takes a bit of upfront work, but if you do it right, it can save you a lot of time in the long run. I've found that this route has worked pretty well for me.... YMMV.