There is a reasonably simple and very effective solution here. The Debian project supports, and in fact encourages co-maintainers for packages. This is a great way to get more manpower into the process and improve the quality of packages. The co-maintainer doesn't even have to be an official member of Debian if the maintainer sponsors the combined work.
I am a Debian developer, and one of the packages that I maintain has been patched by Ubuntu. I only found out about it by looking over the Ubuntu patch site. What I would like to see is the Ubuntu developer contact me, ask to be a co-maintainer, and get those changes directly into the Debian package. This is good for Debian - we get additional help in doing a good job. This is good for Ubuntu since they don't have to re-merge patches every six months. It helps the two groups act as a team, feel good about each other, and save on overall work. And, as the article points out, the increased compatibility between Debian, Ubuntu and all other Debian based distributions (including Knoppix) is a win for end users.
Now that Ubuntu is a rising star, and Debian has just finished Project Lead elections, I would like to see the leadership of the two organizations get together, discuss the idea, and hopefully agree that this is a good way to work together. The leadership can then promote co-maintainership as a 'best practice' within their own organizations, inform the userbase (i.e. get it mentioned on slashdot), PLUS appoint an interoperability liason. The liason's job is to hassle^H^H^H^H^H^H talk with individual developers to help make sure this actually happens. Branden, don't you think this would be a great first accomplishment as DPL?
Of course, there will still be some places where Debian and Ubuntu want to do something differently, so some packages will always be a little incompatible. But the bulk of the 'heavy lifting' across the thousands of packages is all about stuff developers generally agree on. Updating software, finding and fixing problems, improving quality. Ian Murdock is worried an impending 'nightmare'. I think if we can work together well, the upcoming Ubuntu/Debian relationship is going to be software distribution's finest hour.
Maybe the Hitachi folks should stop singing and concentrate on shipping. On January 10 Hitachi announced the 500GB 7K500 drive would ship "first quarter 2005". It is now mid April, where is it?
I have to agree, a true open source Java would be a major improvement. On my platform (Debian) I can install python or php or haskell or gcj trivially, and the package maintainers makes sure any rough edges are smoothed out. The net result is it all just works.
Yesterday, I tried to install Java because a colleague is developing a web application with a Java component in the browser, and I wanted to try it and maybe help him out. That meant installing a Debian helper package. Then manually downloading Java, clicking through some license agreement nobody reads. Running the helper package, typing answers to questions, paging through the same license agreement in the terminal, having to type out "yes" and hoping things work. Not impossible, but at least 10X more work than every other piece software on my system, especially if I want to keep up to date with point releases
And of course it ended up not working - my Firefox still refuses to run Java for some reason. I ended up taking the path of least resistance which was VNC to a Windows computer - a pathetic and painful workaround. Normally I'd file a bug and the package maintainer would smooth out these system integration issues, but Sun's redistribution terms are preventing that. To add insult to injury, I think I remember reading that Sun sued Microsoft for NOT distributing Java at one point, which makes me wonder about corporate schizophrenia. If Java were true open source, I'm confident things would finally get smoothed out and I'd be much more inclined to develop for the platform and trust that it would always be there for me.
Furthermore, there's now a few useful open source programs written in Java that also get hampered. For example, the Lucene search toolkit can't currently be built without Sun's SDK. Therefore, it can't run on the Debian autobuilders and gets relagated to second class citizenship status. As both a proprietary and an open source software developer, my life would be easier, happier and I'd be far more Javaphilic (as opposed to just plain grumpy) working on an open source platform.
>200 lines seems almost too much [blah, blah, blah]
While not explicitly mentioned in the writeup, I also programmed the computer to beep when a piece was successfully processed. That added a line right there.
Can I really take credit for all their work, and call it my own?
Maybe talking about the small size of the application was meant as a complement to the supporting software libraries, rather than an attempt at "taking credit"
If it's precise enough it should be able to put the puzzle together by only looking at the back of the pieces, perhaps with intervention whenever two pieces happen to hash to the same shape.
Error correction codes are an important facet of DataGlyphs when one is trying to store a message. However, in this case there isn't any message stored - every glyphmark is devoted to purely positional information. This is called an address carpet, which is described in more detail in this IEEE article.
Actually, the newer PDF specifications and newer PDF viewers (like the totally excellent xpdf utility, and oh yeah, Adobe acroread 5 and onwards) all support JBIG2 compression. JBIG2 is a token based compression technology giving roughly similar file size and image quality compared to DjVu, but with the advantage that everyone and their uncle can deal with the PDF file format.
So, I recommend scanning to TIFF (or TIFF inside PDF). Even if you don't currently have the encoding softeware, you can convert to JBIG2
compression later as it becomes more and more ubiquitous in the future.
And definitely use a automated document feeder of some sort to keep
from going crazy.
Newer Xerox machines work pretty well for this (I use a DocumentCentre 440ST for this all the time) unless you have hundreds of thousands of pages to deal with, in which case you should either invest in industrial scanning equipment or outsource to a scanning center that does.
A few years and a several billion dollars into the Big Dig, transportation planners discovered a cheaper and easier alternative that would do an even better job easing traffic congestion in Boston and improving the scenic quality of the city. Instead of the "big dig" approach of burying a 8 miles of highway, the state would instead construct a large brick or concrete barrier right across I-93, strong enough to withstand the steady stream of cars crashing into it as they try to get into the city. The section of highway beyond this barrier is then demolished. This approach avoids the "if you build it they will come" increase of cars that quickly negates the effectiveness of most road and highway improvements. Unfortunately, by the time this was realized, enough money had been spent
that it was too difficult to change course.
casual desktop users may want to try Knoppix
on
Debian 3.0r2 Released
·
· Score: 5, Informative
If you are a desktop user and want to see a different perspective of Debian, Knoppix may be a better starting point. Debian 3.0r2 is Debian Stable, which is very good if you prize stability and don't really want your OS to change out from under you. Knoppix is basically a closely tracked derivative of Debian Unstable, and therefore has more recent software which is often desirable for desktop users. For example, the most recent Knoppix ISO was cut on November 19th of this year, so it is REALLY current.
The other nice thing about Knoppix is that it is very easy to try out, and it also makes for a very painless Debian installer. I use it all the time to install Debian Unstable onto x86 desktop machines (see knoppix-installer in/usr/local/bin). I've been a Debian Developer for several years now, and I've pretty much switched over to using Knoppix for all my installation needs.
It's never too late to change your mind. Debian is fairly gung-ho on the idea of multiple maintainers for packages, which makes for a reasonable entry point for folks from other distributions to join the fun. Another good hook is Debian's Testing distribution. The basic idea is to have a usable software collection with a time granularity between Unstable (daily) and Stable (roughly two year cycle). Testing doesn't currently get the tons of programmer nourishment it so richly deserves, and instead relies on automated QA programs and a few gallant human efforts. With some additional manpower, this would be a great layer in the distribution/maintenance system for something like Fedora to emerge.
I installed Knoppix 3.2 this morning on an FIC E-Cube, in all its blue glowing glory. The biggest change to my eye is easier access to Knoppix-specific configuration; that now has its own root menu on the task bar. I also like the new desktop wallpaper which looks like an industrial cave painting. For some reason today's Knoppix didn't see hyperthreaded Pentium 4 as an SMP machine, which Knoppix 3.2 had no problem recognizing. Other than that, no real problems. I went ahead and used Knoppix as installer for Debian - this is definitely my preferred way to install Linux these days.
7 more years of news.com.com.com.com.com?
on
CNET News.com Turns 7
·
· Score: 4, Funny
I worked as an undergrad in the MIT lab in the mid-1990's and probably the first thing I did was break the holovideo system. I had no idea that the tip of a soldering iron is electrically grounded; I touched it to a live wire and promptly sent the power supply up in smoke and very nearly fried some custom wirewrap circuit boards.
Fortunately, I wasn't fired (thanks Steve!)
and I learned a lot about holography over the next few years. While I don't work on holograms today, that early exposure to research was invaluable and convinced me to pursue a career in image processing.
The MIT holovideo system does compute interference patterns, which are used to diffract light. It's the real deal in terms of focusing
light in the right place. A lot of math techniques are used to reduce the computation, but the important part is there - directing light in the right places.
I don't know what's changed over the last half decade or so, but "way back then" there was one main difference between the holovideo system and traditional holograms.
For holovideo, the diffraction patterns were calculated from a whole bunch of 2-D computer graphic images (i.e. the view from each angle) rather than a real live 3-D object.
Perceptually, there is no significant difference between a holographic stereogram and a hologram, as long as enough viewing angles are used. But from a technical standpoint the creation technique is different -- so it has a different name.
By the way, one of the biggest annoyances was showing off the state-of-the-art holovideo system or still holograms to visitors, and having people consistantly say "Wow, those holograms look really bad." Everyone just assumed we'd at least be as good as Princess Leia in Star Wars; after all that movie was made decades ago, right?
I grew up in Vermont. Vermont is a place where
billboards are illegal and businesses are resticted
to a single sign of limited size on their building.
It's a place where quality of life (in this case
preserving some scenic beauty) is valued more
than the short sighted whims of businesses. Makes me proud -- and is a huge contrast to what I see living in the San Francisco Bay Area
where worshipping industry seems to be
people's number one hobby.
And for those who feel pesky quality-of-life laws and regulations cost business
too much money, I'm sure New Jersey can accomodate.
Does the govt really think that crypto export restrictions have prevented terrorists from having strong crypto?
No.
The government thinks that they can discourage use of cryptography by the general public. Thus those using cryptography will stand out, and thus can be given greater scrutiny. There are much easier ways
to find the contents of a message than breaking cryptography (for example, keystroke recorders).
Start simple. Try something like multimouse, which provides
a shared pointer (a little triangular marker)
for images on a web page. It's a trivially
simple java application -- see the FICT project page.
You'd be surprised how much this alone can buy you.
Especially compared to shared whiteboards with
artistically challenged people trying to draw
with their mouse.
On a more serious note, how does Xerox deal with large white spaces? How do you keep the rows lined up? There doesn't appear to be much room for error. What happens when someone spills coffee (or even water) on the document? How much redundency is built in?
Actually, we're quite robust. Read the article for details about synchronization in DataGlyphs.
Redundancy is adjustable, and I can personally
attest that DataGlyphs have survived my coffee spill. Haven't tried Guiness yet.
DataGlyphs are flexible in that they are robust, have a high data density, and can be cosmetically pleasing. Traditionally people consider (some) barcodes good for the former and (some) watermarks good for the latter.
Slashdot Mirror
There is a reasonably simple and very effective solution here. The Debian project supports, and in fact encourages co-maintainers for packages. This is a great way to get more manpower into the process and improve the quality of packages. The co-maintainer doesn't even have to be an official member of Debian if the maintainer sponsors the combined work.
I am a Debian developer, and one of the packages that I maintain has been patched by Ubuntu. I only found out about it by looking over the Ubuntu patch site. What I would like to see is the Ubuntu developer contact me, ask to be a co-maintainer, and get those changes directly into the Debian package. This is good for Debian - we get additional help in doing a good job. This is good for Ubuntu since they don't have to re-merge patches every six months. It helps the two groups act as a team, feel good about each other, and save on overall work. And, as the article points out, the increased compatibility between Debian, Ubuntu and all other Debian based distributions (including Knoppix) is a win for end users.
Now that Ubuntu is a rising star, and Debian has just finished Project Lead elections, I would like to see the leadership of the two organizations get together, discuss the idea, and hopefully agree that this is a good way to work together. The leadership can then promote co-maintainership as a 'best practice' within their own organizations, inform the userbase (i.e. get it mentioned on slashdot), PLUS appoint an interoperability liason. The liason's job is to hassle^H^H^H^H^H^H talk with individual developers to help make sure this actually happens. Branden, don't you think this would be a great first accomplishment as DPL?
Of course, there will still be some places where Debian and Ubuntu want to do something differently, so some packages will always be a little incompatible. But the bulk of the 'heavy lifting' across the thousands of packages is all about stuff developers generally agree on. Updating software, finding and fixing problems, improving quality. Ian Murdock is worried an impending 'nightmare'. I think if we can work together well, the upcoming Ubuntu/Debian relationship is going to be software distribution's finest hour.
Maybe the Hitachi folks should stop singing and concentrate on shipping. On January 10 Hitachi announced the 500GB 7K500 drive would ship "first quarter 2005". It is now mid April, where is it?
I have to agree, a true open source Java would be a major improvement. On my platform (Debian) I can install python or php or haskell or gcj trivially, and the package maintainers makes sure any rough edges are smoothed out. The net result is it all just works.
Yesterday, I tried to install Java because a colleague is developing a web application with a Java component in the browser, and I wanted to try it and maybe help him out. That meant installing a Debian helper package. Then manually downloading Java, clicking through some license agreement nobody reads. Running the helper package, typing answers to questions, paging through the same license agreement in the terminal, having to type out "yes" and hoping things work. Not impossible, but at least 10X more work than every other piece software on my system, especially if I want to keep up to date with point releases
And of course it ended up not working - my Firefox still refuses to run Java for some reason. I ended up taking the path of least resistance which was VNC to a Windows computer - a pathetic and painful workaround. Normally I'd file a bug and the package maintainer would smooth out these system integration issues, but Sun's redistribution terms are preventing that. To add insult to injury, I think I remember reading that Sun sued Microsoft for NOT distributing Java at one point, which makes me wonder about corporate schizophrenia. If Java were true open source, I'm confident things would finally get smoothed out and I'd be much more inclined to develop for the platform and trust that it would always be there for me.
Furthermore, there's now a few useful open source programs written in Java that also get hampered. For example, the Lucene search toolkit can't currently be built without Sun's SDK. Therefore, it can't run on the Debian autobuilders and gets relagated to second class citizenship status. As both a proprietary and an open source software developer, my life would be easier, happier and I'd be far more Javaphilic (as opposed to just plain grumpy) working on an open source platform.
While not explicitly mentioned in the writeup, I also programmed the computer to beep when a piece was successfully processed. That added a line right there.
Maybe talking about the small size of the application was meant as a complement to the supporting software libraries, rather than an attempt at "taking credit"
See A Global Approach to Automatic Solution of Jigsaw Puzzles
Error correction codes are an important facet of DataGlyphs when one is trying to store a message. However, in this case there isn't any message stored - every glyphmark is devoted to purely positional information. This is called an address carpet, which is described in more detail in this
IEEE article.
PARC (and others) have already tackled that problem. Here's my favorite research paper on the topic.
Goldberg, D.; Malon, C.; Bern, M. W. A global approach to automatic solution of jigsaw puzzles. Computational Geometry. 2004 June; 28 (2): 165-174.
Yes, but is it running Electric Grill Linux?
I agree.
So, I recommend scanning to TIFF (or TIFF inside PDF). Even if you don't currently have the encoding softeware, you can convert to JBIG2 compression later as it becomes more and more ubiquitous in the future.
And definitely use a automated document feeder of some sort to keep from going crazy. Newer Xerox machines work pretty well for this (I use a DocumentCentre 440ST for this all the time) unless you have hundreds of thousands of pages to deal with, in which case you should either invest in industrial scanning equipment or outsource to a scanning center that does.
A few years and a several billion dollars into the Big Dig, transportation planners discovered a cheaper and easier alternative that would do an even better job easing traffic congestion in Boston and improving the scenic quality of the city. Instead of the "big dig" approach of burying a 8 miles of highway, the state would instead construct a large brick or concrete barrier right across I-93, strong enough to withstand the steady stream of cars crashing into it as they try to get into the city. The section of highway beyond this barrier is then demolished. This approach avoids the "if you build it they will come" increase of cars that quickly negates the effectiveness of most road and highway improvements. Unfortunately, by the time this was realized, enough money had been spent that it was too difficult to change course.
The other nice thing about Knoppix is that it is very easy to try out, and it also makes for a very painless Debian installer. I use it all the time to install Debian Unstable onto x86 desktop machines (see knoppix-installer in /usr/local/bin). I've been a Debian Developer for several years now, and I've pretty much switched over to using Knoppix for all my installation needs.
It's never too late to change your mind. Debian is fairly gung-ho on the idea of multiple maintainers for packages, which makes for a reasonable entry point for folks from other distributions to join the fun. Another good hook is Debian's Testing distribution. The basic idea is to have a usable software collection with a time granularity between Unstable (daily) and Stable (roughly two year cycle). Testing doesn't currently get the tons of programmer nourishment it so richly deserves, and instead relies on automated QA programs and a few gallant human efforts. With some additional manpower, this would be a great layer in the distribution/maintenance system for something like Fedora to emerge.
Push F2 when you boot Knoppix (when it says "push F2 for help") and it will tell you.
I installed Knoppix 3.2 this morning on an FIC E-Cube, in all its blue glowing glory. The biggest change to my eye is easier access to Knoppix-specific configuration; that now has its own root menu on the task bar. I also like the new desktop wallpaper which looks like an industrial cave painting. For some reason today's Knoppix didn't see hyperthreaded Pentium 4 as an SMP machine, which Knoppix 3.2 had no problem recognizing. Other than that, no real problems. I went ahead and used Knoppix as installer for Debian - this is definitely my preferred way to install Linux these days.
What, did someone break a mirror, AGAIN?
FlowPort is a Xerox product involving copiers that came out of PARC - it also uses DataGlyphs, by the way.
The MIT holovideo system does compute interference patterns, which are used to diffract light. It's the real deal in terms of focusing light in the right place. A lot of math techniques are used to reduce the computation, but the important part is there - directing light in the right places.
I don't know what's changed over the last half decade or so, but "way back then" there was one main difference between the holovideo system and traditional holograms. For holovideo, the diffraction patterns were calculated from a whole bunch of 2-D computer graphic images (i.e. the view from each angle) rather than a real live 3-D object. Perceptually, there is no significant difference between a holographic stereogram and a hologram, as long as enough viewing angles are used. But from a technical standpoint the creation technique is different -- so it has a different name.
By the way, one of the biggest annoyances was showing off the state-of-the-art holovideo system or still holograms to visitors, and having people consistantly say "Wow, those holograms look really bad." Everyone just assumed we'd at least be as good as Princess Leia in Star Wars; after all that movie was made decades ago, right?
And for those who feel pesky quality-of-life laws and regulations cost business too much money, I'm sure New Jersey can accomodate.
No.
The government thinks that they can discourage use of cryptography by the general public. Thus those using cryptography will stand out, and thus can be given greater scrutiny. There are much easier ways to find the contents of a message than breaking cryptography (for example, keystroke recorders).
Start simple. Try something like multimouse, which provides a shared pointer (a little triangular marker) for images on a web page. It's a trivially simple java application -- see the FICT project page. You'd be surprised how much this alone can buy you. Especially compared to shared whiteboards with artistically challenged people trying to draw with their mouse.
For starters, "Printed Embedded Data Graphical
User Interfaces." Read the article.
Actually, we're quite robust. Read the article for details about synchronization in DataGlyphs. Redundancy is adjustable, and I can personally attest that DataGlyphs have survived my coffee spill. Haven't tried Guiness yet.
DataGlyphs are flexible in that they are robust, have a high data density, and can be cosmetically pleasing. Traditionally people consider (some) barcodes good for the former and (some) watermarks good for the latter.