Does anyone know of a robot (built with Mindstorm product or some other) that can flip a book and turn its pages as it photocopies it on a standard home scanner?
This would help me in my book digitization project.
For around $150, you get a couple of megabits/sec streamed to your ears.
Unfortunately I don't know how quality the supplied headphones are, although it might not be too hard to hack up your headphones by ripping the receiver out of the Amphony headset and warming up the soldering iron.
Of course for much more cash, buy an iPAQ and plug in an 802.11b card:)
Sorry, you couldn't be more wrong. Richard Fateman of the Berkeley CS department tried to write in LISP a computer algebra package that could parse what you call Mathematica's "open" syntax.
Wolfram threatened legal action against Fateman and ultimately forced him to abandon his project.
Re. MathML, I'm not surprised that Wolfram helped draft it. After staring at the grossly verbose syntax for a few minutes, you realize how superior Mathematica's native syntax is (again proving what a genius Wolfram was to come up with it in the first place). But don't fall too in love with it, because if you try to write your own WYSIWYG editor to parse (god forbid actually evaluate) native Mathematica syntax, you can also expect a letter from Wolfram's legal department.
I personally prefer a slick interface based on a standard like TeX and Metafont, which is why I again urge you to visit www.TeXmacs.org:)
This is what Wolfram should have attempted, but oops, if they based it on the already freely available work of Donald Knuth instead of reinventing the wheel themselves (or did they, I guess we'll never know), maybe folks would be less willing to pay over a grand for it.
REDUCE is not some "hacked up piece of freeware". Visit its homepage and use the free trial before you shoot from the hip next time.
And concerning using Mathematica "to get things done", I'd advise journals not to accept for publication papers that rely on calculations done in closed source computer algebra packages.
One of the general ideals in publishing academic research is full disclosure so others can verify or disprove your claims. Kind of hard to do when certain pieces of the application are closed black boxes (for example the Simplify command). In fact, it was precisely because of the opaque nature of Simplify (which constantly changed from version to version) in not just Mathematica but Maple and others that prompted one group of physicists to start their own project using C++. Now I'm no fan of C++ (prefer the Lisp-inspired syntax of Mathematica any day) but I think this example does a lot to prove my case. The name of this project is Ginac
The book sounds like a belated promo for cellular automata (that stuff was cool a few decades ago when Conway at Princeton was playing around with the game of Life on his computer... Conway is a real mathematician by the way who has done some very legitimate work.)
The real genius of Wolfram is not his "formula" that claims to explain the world, but how he has bilked universities and research institutes around the world in the untold millions for Mathematica site licenses.
It is not an understatement to call Wolfram the Microsoft of scientific computing software. Both Gates and Wolfram dropped out of academia to create their respective computing empires. Both use proprietary data formats to lock in their customers. Both go to school campuses and offer students the "first hit" for free.
Gates' used a little of Dad's money (wealthy Washington lawyer) to get his start. Wolfram, in a stroke of true genius, used his McArthur grant
to set up shop.
The book is just a farce to make the hungry PhD Computer Science students who bang their heads trying to fix the bugs in Mathematica (and there are many) feel like Wolfram is doing something useful for his hefty paycheck while they sweat over their mundane chores. Can't you just hear them whispering to each other "When is Stevie Wonderboy going to tell us how the universe got started?"
If you are at a university that has a site license for Mathematica, ask the university to consider canceling the license and purchasing the open-source REDUCE system instead. It is an older product than Mathematica and lacks a slick GUI interface. This is no longer a problem though because REDUCE interfaces nicely with TeXmacs, and if you haven't heard about the latter, check out this Metafont-based WYSIWYG scientific editor at www.texmacs.org.
Well I couldn't agree more about lugging around dead trees... I've got over 200 mathematics textbooks myself that I am in the process of digitizing.
Having done a fair amount of research on the topic, this is what I can share:
If you are willing to feed your books to the guillotine, then it should be enough to purchase a home/office level scanner that includes a automatic document feeder (ADF). My advice here is to buy an Epson product. Even their lowest end models are lightning fast for their class, and the quality of the scan is consistently high.
For more serious scanning of texts that you don't want to destroy I settled on the Ricoh IS450SE, because this model has a 11"x17" glass plate, which is ideal for scanning both pages at the same time. It copies two pages in just over one second, which makes quick work of manually scanning a 500 page textbook. It's even quiet enough to operate while watching your favorite TV program. The native resolution is 400dpi and it can handle grayscale, at a slightly slower speed. The scanner was my top choice because of speed and cost. At $3000 street price (including ADF), I could justify it because I'll make that back selling my textbooks used on the net, not to mention avoiding purchasing any new textbooks. Some of the other competing manufacturers in this category include Fujitsu, Panasonic, Kodak, and Xerox. Visioneer has recently introduced a sub $1000 scanner that is supposed to be competitive in this category of machines which might be worth investigating.
One last comment about hardware. Since my goal is to read my texts online, I purchased a laptop with a 15" 1600x1200 screen from Dell which I highly recommend. Apparently Dell has recently introduced an improved version of this UXGA screen called "UltraSharp" which supposedly fixes some problems with uniformity of contrast from top to bottom of screen, which might be of interest to someone considering purchasing a 133dpi LCD laptop. (IBM, Compaq, Sony and Hitachi I believe also offer models with UXGA screens.) I'm curious if anyone knows the current lightweight champion for laptops with UXGA screens. My Dell is over 9 lbs:)
When it comes to software, I had to write my own Python script using the freely downloadable TWAIN interface kit for Python since the driver interface to various software packages I tried was not optimized for doing bulk scanning. In particular, I wanted the software to automatically name/number the images as they became available. One of the Ricoh TWAIN drivers I tried was a little buggy but I was able to correct for that in the Python script I wrote by seeking out the TWAIN spec from www.twain.org and reading up on it. I can make my mods available to anyone interested in driving a Ricoh product from Ricoh's supplied TWAIN drivers.
For archiving/image processing of the raw scanned images there are two formats I've looked carefully at using. The first is DjVu available from djvu.sourceforge.net. This uses the open standard called jbg2 to efficiently compress images of scanned text (by indexing recognized glyphs). It can also automatically segment images according to text and image and compress the latter with wavelet technology. AT&T spun off a commercial venture called LizardTech that is based on the same software provided on sourceforge albeit a little faster and more sophisticated. The open source tools should be more than adequate however for non-commercial applications.
There is also of course PDF, the latest version of which supports jbg2 encoding. Unfortunately the linux viewers for pdf do not support this feature in the latest PDF spec, and neither the PDF viewers or the DjVu viewers support sub-pixel rendering of images. (Of course Adobe Acrobat 5 does so through its "CoolType" technology, but that is only available on Windows platform as far as I know.) I am currently working to upgrade the opensource DjVu viewer to support "pixel borrowing" on LCD screens since standard grayscale anti-aliasing (supported by the viewer) does not look so great on LCD screens.
Then there is the open source OCR project called claraocr. My ultimate goal is to add support to this package for recognizing mathematical formulas and producing equivalent TeX code. This is a HARD problem and I don't expect to make significant progress on it anytime soon.
Anyway, I hope the above the long-winded tour through my adventures in this area can provide some useful insights.
Good luck, and make sure to make available your handiwork on P2P networks like gnutella and Freenet:)
In many ways modern day universities are still stuck in the Middle Ages.
Their elaborate hierarchies in the administration (chancellor, dean, provost, etc.) with titles more fitting of the British royalty, the infamous 5-10 year graduate student advisor-advisee relationship that reeks of indentured servitude, and the undergraduate Displinary Boards that are run like Inquisitions,... are all but a few reasons why.
Anyone know of any schools that treat their students more like customers than like children?
I'm generally interested in concrete areas where the GNOME project aims to innovate the GUI (as opposed to merely clone the "competition", be it KDE, Windows, or whatever).
What can we expect from GNOME that doesn't exist or have a planned counterpart in other GUIs? Sure, open source is a powerful concept (I've recently started to use some source code to tweak functionality according to my preferences), but at the level of the GUI user experience, what features are there that could be reasonably considered "uniquely GNOME". (I'm asking at the level of the user, not the developer, as I'm already aware of many distinctions that can be made at this level.)
As a somewhat relevant example to the above question, consider the OpenOffice project. It is great to have a open source project that mimics much of the capability found in Microsoft Office. OTOH, there is an WYSIWYG editor called TeXmacs that is remarkably innovative and solves many problems that scientific writers have with traditional Word-style word processors.
I'd personally like to see TeXmacs (now that a stable version has been released) be adopted whole-heartedly by the GNOME project, not necessarily to replace OpenOffice, but because this fine piece of software deserves the extra attention it could attract from being an official part of the GNOME desktop suite of apps. Of course there would be some work required to switch over from the XForms toolkit to GDK, but this is exactly the kind of software that could shift GNOME toward the avant-gard, and away from the image it sometimes projects as trying to "play catch-up" to other GUIs out there.
IMHO, TeXmacs is a superior piece of software in every regard.
Most importantly, it is truly What-You-See-Is-What-You-Get, where as Lyx is only What-You-See-Is-What-You-Mean. *Both* screen and paper output from TeXmacs is based on the highly-respected Metafont font rasterization engine.
TeXmacs improves on traditional TeX typesetting not just in providing a real-time rendering engine but also in some more subtle ways, like how it handles various typesetting issues (including font-switching) semantically. Traditional TeX never provided an easy interface for users to switch fonts, although LaTeX eventually improved the situation considerably.
TeXmacs even surpasses commerical offerings like BlueSky's Textures, which renders the.DVI output in real time, but in a separate window which is not editable (you always have to return to the TeX source code to make changes.) And Textures is not available for Linux:)
I think that TeXmacs provides a nice counterpoint to development efforts like OpenOffice and KWord, which appear to want to clone Microsoft Word as closely as possible. TeXmacs on the other hand is a real innovation for technical word processing, and I think deserves at least as much attention from the development and user community as the other more office-oriented suites. (In particular, I'd suggest folks working on KFormula to abandon their efforts and see if they can integrate TeXmacs as KPart into KOffice !!!)
Congrats to Slashdot for recognizing the hardwork of the TeXmacs team to produce the 1.0 release.
Future additions to TeXmacs that I hope really come to pass (and to which I might even lend my own efforts time permitting) include:
1. integrated sketching facility based on SVG... say a nice GUI to the functionality of a traditional TeX package like PSTricks which uses Postscript commands to draw on top of TeX output.
2. sub-pixel anti-aliasing for LCD screens
3. adoption of a more standard toolkit like GTK or Qt
4. better documentation of the source since an important project like this needs all the developers it can get (David Allouche's efforts in this regard are a great start)
5. better integration with computer algebra software... in particular ability to cut and paste between CAS output and input prompts cleanly and consistently
TeXmacs is one of the most exciting pieces of open-source software I've seen in a long time.
Do not fail to at least give it a spin if technical typesetting is at all a part of your workday.
It appears that the "wizardry" theme for naming tools in the Sorceror distro has put off more than a few people.
Casting "spells" is appropriate however given the name of the distro, which itself is a clever choice given the binary-free nature of the installation process.
Perhaps they should have called it Sourcerer Linux:)
A recent PhD graduate trying to figure out whether or not to remain in academia runs the following numbers:
Teaching: roughly 25 weeks of work at 10 hours per week for $50K = almost $200/hr.
Management Consulting: roughly 50 weeks of work at 50 hours per week for $120K = almost $50/hr.
(Of course some of the money for the "postdoc" teaching job is for research, but I don't include that because many PhDs enjoy their research and would pursue it as a hobby regardless of its utility to society without getting paid a dime if they were independently wealthy.)
My question is whether or not others believe this OpenCourse initiative (at MIT and elsewhere) will ultimately spoil the party for academics.
(Clearly for many the concern is already there... OpenCourseWare takes great pains in their literature to assure other MIT faculty that the course websites will do nothing to steal away the magic of an MIT education that can only be found on the MIT campus proper.)
Does anyone have any insight into why Blender is not open-source?
I'm wary about investing time in learning a 3D-modeling environment that is free today, but could easily cost more money than I can afford tomorrow.
Does anyone know if NaN plans to truly open-source Blender at some point in the future?
Also can anyone compare Blender to open-source programs like Panorama and K3D?
Slashdot Mirror
Does anyone know of a robot (built with Mindstorm
product or some other) that can flip a book and
turn its pages as it photocopies it on a standard
home scanner?
This would help me in my book digitization
project.
Check out Amphony.com.
:)
For around $150, you get a couple of megabits/sec
streamed to your ears.
Unfortunately I don't know how quality the
supplied headphones are, although it might not
be too hard to hack up your headphones by ripping
the receiver out of the Amphony headset and
warming up the soldering iron.
Of course for much more cash, buy an iPAQ and
plug in an 802.11b card
Sorry, you couldn't be more wrong. Richard
Fateman of the Berkeley CS department tried to
write in LISP a computer algebra package that
could parse what you call Mathematica's
"open" syntax.
Wolfram threatened legal action against Fateman
and ultimately forced him to abandon his project.
Re. MathML, I'm not surprised that Wolfram helped
draft it. After staring at the grossly verbose
syntax for a few minutes, you realize how
superior Mathematica's native syntax is (again
proving what a genius Wolfram was to come up
with it in the first place). But don't fall
too in love with it, because if you try to write
your own WYSIWYG editor to parse (god forbid
actually evaluate) native Mathematica syntax,
you can also expect a letter from Wolfram's
legal department.
I personally prefer a slick interface based on a :)
standard like TeX and Metafont, which is why
I again urge you to visit www.TeXmacs.org
This is what Wolfram should have attempted, but
oops, if they based it on the already freely
available work of Donald Knuth instead of
reinventing the wheel themselves (or did they,
I guess we'll never know), maybe folks would
be less willing to pay over a grand for it.
REDUCE is not some "hacked up piece of freeware".
Visit its homepage and use the free trial before
you shoot from the hip next time.
And concerning using Mathematica "to get things
done", I'd advise journals not to accept for
publication papers that rely on calculations done
in closed source computer algebra packages.
One of the general ideals in publishing academic
research is full disclosure so others can verify
or disprove your claims. Kind of hard to do
when certain pieces of the application are
closed black boxes (for example the Simplify
command). In fact, it was precisely because of
the opaque nature of Simplify (which constantly
changed from version to version) in not just
Mathematica but Maple and others that prompted
one group of physicists to start their own
project using C++. Now I'm no fan of C++ (prefer
the Lisp-inspired syntax of Mathematica any day)
but I think this example does a lot to prove my
case. The name of this project is Ginac
The book sounds like a belated promo for cellular automata (that stuff was cool a few decades ago when Conway at Princeton was playing around with the game of Life on his computer ... Conway is a real mathematician by the way who has done some very legitimate work.)
The real genius of Wolfram is not his "formula" that claims to explain the world, but how he has bilked universities and research institutes around the world in the untold millions for Mathematica site licenses.
It is not an understatement to call Wolfram the Microsoft of scientific computing software. Both Gates and Wolfram dropped out of academia to create their respective computing empires. Both use proprietary data formats to lock in their customers. Both go to school campuses and offer students the "first hit" for free.
Gates' used a little of Dad's money (wealthy Washington lawyer) to get his start. Wolfram, in a stroke of true genius, used his McArthur grant to set up shop.
The book is just a farce to make the hungry PhD Computer Science students who bang their heads trying to fix the bugs in Mathematica (and there are many) feel like Wolfram is doing something useful for his hefty paycheck while they sweat over their mundane chores. Can't you just hear them whispering to each other "When is Stevie Wonderboy going to tell us how the universe got started?"
If you are at a university that has a site license for Mathematica, ask the university to consider canceling the license and purchasing the open-source REDUCE system instead. It is an older product than Mathematica and lacks a slick GUI interface. This is no longer a problem though because REDUCE interfaces nicely with TeXmacs, and if you haven't heard about the latter, check out this Metafont-based WYSIWYG scientific editor at www.texmacs.org.
Just my 2 cents worth; done ranting :)
Guess there is no use offering the rebate to :)
current Debian users since their package
management system beats Red Hat hands down
(Just had to add my 2 cents since I recently
switched back to Debian after trying Gentoo
and not being too impressed by Portage.)
Well I couldn't agree more about lugging around dead trees ... I've got over 200 mathematics textbooks myself that I am in the process of digitizing.
:)
:)
Having done a fair amount of research on the topic, this is what I can share:
If you are willing to feed your books to the guillotine, then it should be enough to purchase a home/office level scanner that includes a automatic document feeder (ADF). My advice here is to buy an Epson product. Even their lowest end models are lightning fast for their class, and the quality of the scan is consistently high.
For more serious scanning of texts that you don't want to destroy I settled on the Ricoh IS450SE, because this model has a 11"x17" glass plate, which is ideal for scanning both pages at the same time. It copies two pages in just over one second, which makes quick work of manually scanning a 500 page textbook. It's even quiet enough to operate while watching your favorite TV program. The native resolution is 400dpi and it can handle grayscale, at a slightly slower speed. The scanner was my top choice because of speed and cost. At $3000 street price (including ADF), I could justify it because I'll make that back selling my textbooks used on the net, not to mention avoiding purchasing any new textbooks. Some of the other competing manufacturers in this category include Fujitsu, Panasonic, Kodak, and Xerox. Visioneer has recently introduced a sub $1000 scanner that is supposed to be competitive in this category of machines which might be worth investigating.
One last comment about hardware. Since my goal is to read my texts online, I purchased a laptop with a 15" 1600x1200 screen from Dell which I highly recommend. Apparently Dell has recently introduced an improved version of this UXGA screen called "UltraSharp" which supposedly fixes some problems with uniformity of contrast from top to bottom of screen, which might be of interest to someone considering purchasing a 133dpi LCD laptop. (IBM, Compaq, Sony and Hitachi I believe also offer models with UXGA screens.) I'm curious if anyone knows the current lightweight champion for laptops with UXGA screens. My Dell is over 9 lbs
When it comes to software, I had to write my own Python script using the freely downloadable TWAIN interface kit for Python since the driver interface to various software packages I tried was not optimized for doing bulk scanning. In particular, I wanted the software to automatically name/number the images as they became available. One of the Ricoh TWAIN drivers I tried was a little buggy but I was able to correct for that in the Python script I wrote by seeking out the TWAIN spec from www.twain.org and reading up on it. I can make my mods available to anyone interested in driving a Ricoh product from Ricoh's supplied TWAIN drivers.
For archiving/image processing of the raw scanned images there are two formats I've looked carefully at using. The first is DjVu available from djvu.sourceforge.net. This uses the open standard called jbg2 to efficiently compress images of scanned text (by indexing recognized glyphs). It can also automatically segment images according to text and image and compress the latter with wavelet technology. AT&T spun off a commercial venture called LizardTech that is based on the same software provided on sourceforge albeit a little faster and more sophisticated. The open source tools should be more than adequate however for non-commercial applications.
There is also of course PDF, the latest version of which supports jbg2 encoding. Unfortunately the linux viewers for pdf do not support this feature in the latest PDF spec, and neither the PDF viewers or the DjVu viewers support sub-pixel rendering of images. (Of course Adobe Acrobat 5 does so through its "CoolType" technology, but that is only available on Windows platform as far as I know.) I am currently working to upgrade the opensource DjVu viewer to support "pixel borrowing" on LCD screens since standard grayscale anti-aliasing (supported by the viewer) does not look so great on LCD screens.
Then there is the open source OCR project called claraocr. My ultimate goal is to add support to this package for recognizing mathematical formulas and producing equivalent TeX code. This is a HARD problem and I don't expect to make significant progress on it anytime soon.
Anyway, I hope the above the long-winded tour through my adventures in this area can provide some useful insights.
Good luck, and make sure to make available your handiwork on P2P networks like gnutella and Freenet
In many ways modern day universities are still
... are all but a few
stuck in the Middle Ages.
Their elaborate hierarchies in the administration
(chancellor, dean, provost, etc.) with titles
more fitting of the British royalty, the infamous
5-10 year graduate student advisor-advisee
relationship that reeks of indentured servitude,
and the undergraduate Displinary Boards that are
run like Inquisitions,
reasons why.
Anyone know of any schools that treat their
students more like customers than like children?
I'm generally interested in concrete areas where
the GNOME project aims to innovate the GUI (as
opposed to merely clone the "competition", be
it KDE, Windows, or whatever).
What can we expect from GNOME that doesn't
exist or have a planned counterpart in other GUIs?
Sure, open source is a powerful concept
(I've recently started to use some source code
to tweak functionality according to my preferences),
but at the level of the GUI user experience, what
features are there that could be reasonably
considered "uniquely GNOME". (I'm asking at the
level of the user, not the developer, as I'm
already aware of many distinctions that can be
made at this level.)
As a somewhat relevant example to the above
question, consider the OpenOffice project.
It is great to have a open source project that
mimics much of the capability found in Microsoft
Office. OTOH, there is an WYSIWYG editor called
TeXmacs that is remarkably innovative and solves
many problems that scientific writers have with
traditional Word-style word processors.
I'd personally like to see TeXmacs (now that a
stable version has been released) be adopted
whole-heartedly by the GNOME project, not
necessarily to replace OpenOffice, but because
this fine piece of software deserves the extra
attention it could attract from being an official
part of the GNOME desktop suite of apps. Of
course there would be some work required to
switch over from the XForms toolkit to GDK, but
this is exactly the kind of software that could
shift GNOME toward the avant-gard, and away from
the image it sometimes projects as trying to "play
catch-up" to other GUIs out there.
You state that Axiom is now being released
as open source.
Where can one download the sources?
IMHO, TeXmacs is a superior piece of software in every regard.
Most importantly, it is truly What-You-See-Is-What-You-Get, where as Lyx is only What-You-See-Is-What-You-Mean. *Both* screen and paper output from TeXmacs is based on the highly-respected Metafont font rasterization engine.
TeXmacs improves on traditional TeX typesetting not just in providing a real-time rendering engine but also in some more subtle ways, like how it handles various typesetting issues (including font-switching) semantically. Traditional TeX never provided an easy interface for users to switch fonts, although LaTeX eventually improved the situation considerably.
TeXmacs even surpasses commerical offerings like BlueSky's Textures, which renders the
I think that TeXmacs provides a nice counterpoint to development efforts like OpenOffice and KWord, which appear to want to clone Microsoft Word as closely as possible. TeXmacs on the other hand is a real innovation for technical word processing, and I think deserves at least as much attention from the development and user community as the other more office-oriented suites. (In particular, I'd suggest folks working on KFormula to abandon their efforts and see if they can integrate TeXmacs as KPart into KOffice !!!)
Congrats to Slashdot for recognizing the hardwork of the TeXmacs team to produce the 1.0 release.
Future additions to TeXmacs that I hope really come to pass (and to which I might even lend my own efforts time permitting) include:
1. integrated sketching facility based on SVG
2. sub-pixel anti-aliasing for LCD screens
3. adoption of a more standard toolkit like GTK or Qt
4. better documentation of the source since an important project like this needs all the developers it can get (David Allouche's efforts in this regard are a great start)
5. better integration with computer algebra software
TeXmacs is one of the most exciting pieces of open-source software I've seen in a long time.
Do not fail to at least give it a spin if technical typesetting is at all a part of your workday.
It appears that the "wizardry" theme for naming
:)
tools in the Sorceror distro has put off more
than a few people.
Casting "spells" is appropriate however given
the name of the distro, which itself is a clever
choice given the binary-free nature of
the installation process.
Perhaps they should have called it Sourcerer Linux
A recent PhD graduate trying to figure out
...
whether or not to remain in academia runs the
following numbers:
Teaching: roughly 25 weeks of work at 10 hours
per week for $50K = almost $200/hr.
Management Consulting: roughly 50 weeks of work
at 50 hours per week for $120K = almost $50/hr.
(Of course some of the money for the "postdoc"
teaching job is for research, but I don't include
that because many PhDs enjoy their research and
would pursue it as a hobby regardless of its
utility to society without getting paid
a dime if they were independently wealthy.)
My question is whether or not others believe this
OpenCourse initiative (at MIT and elsewhere) will
ultimately spoil the party for academics.
(Clearly for many the concern is already there
OpenCourseWare takes great pains in their
literature to assure other MIT faculty that
the course websites will do nothing to steal away
the magic of an MIT education that can only be
found on the MIT campus proper.)
Does anyone have any insight into why Blender is not open-source? I'm wary about investing time in learning a 3D-modeling environment that is free today, but could easily cost more money than I can afford tomorrow. Does anyone know if NaN plans to truly open-source Blender at some point in the future? Also can anyone compare Blender to open-source programs like Panorama and K3D?