16 bit typically refers to 16 bits per what I'll call a sample. 24 bit RGB has 8 bits per sample, 3 samples per pixel. 16 bit samples would be 16x3 = 48 bit RGB.
You'd have to design it with "instant-retaliation". Every time you mark a driver, the driver's score goes up one point, but your score also goes up one point. This would discourage people from blindly marking everyone they see. Only cars that had been marked [hopefully] independently would exceed the threshold for a ticket. This system would still be vulnerable to a coordinated attack where a band of trolls each mark the same target. This might be mitigated somewhat by changing the 1-1 retaliatino ratio.
You could also have a meta moderation that was simply based on the rate at which you doled out marks. If you exceed a certain rate (say, two standard deviations above average), then your marks are ignored.
Re:Lisp not accessible?
on
Beyond Java
·
· Score: 1
Emacs may well lack some features... but for anti-aliased text one can
always run it from an appropriate X terminal emulator. I haven't used
it, but there is a version that uses Gtk2 and presumably has anti-aliased
text.
Agreed. I think the lawyer's job is not to argue to get "bad guys" off the
hook (and perhaps it isn't; I don't see lawyers outside of TV or movies), but
to ensure a fair trial.
Given that the Toyota is marginally more fuel efficient than a small gas or
diesel car, the only good from buying one seems to be supporting hybrid
research with your dollar (which is a good thing). Fuel efficiency folks also
neglect the energy required to make the car in the first place. A used (the
manufacturing energy is a sunk cost) but fuel efficient car probably wins in
total energy. Of course, that's more of a short term thing; not everyone can
buy used cars forever. I have no idea how short "short term" means in this
case. Newer cars also tend to pollute less due to better catalytic converters,
but a two to five year old car will probably be about the same in that
department.
In conclusion, I applaud your support of hybrid research. Ride a bike if
you can. It's a shame our (U.S.) culture seems so derisive of doing that, and
that most cities (at least the ones I've seen) are designed exclusively for
cars.
The purpose of the GPL is to ensure that a software program includes the four freedoms. If a
program will only function with, say, a signed binary module, and it is
impossible for the end use to create such a signed module of his own, then he
is missing the "freedom to study how the program works, and adapt it to [his]
needs". I don't really see how adding a new restriction to the GPL goes
against this spirit of the GPL. It may go against your idea of what a license
should be, but to say GPLv3 is a political statement while GPLv2 is not is
wrong because the GPLv3 additions are within the spirit of the four
freedoms.
Can I sell used digital tracks? Can I lend digital tracks to others
deleting my copy while it's lent out? Consider a "hot-potato" or collaborative
library style p2p system where each member contributes the songs he owns to the
pool. When another member wishes to hear a song, she determines which member
or members own the song, borrows (or buys for free) the song (which the owner
then deletes), listens to the song, and then either returns it to the owner or
registers with the pool as the new owner ready to lend to others.
I wonder what a court would think of this... One would most likely need to
come up with a file transfer protocol that could "move" a file rather than
"copy" a file (like, transfer bytes 1 to X, wait for confirmation, delete bytes
1 to X, transfer bytes X to 2*X,...).
Your comment made me think of a situation that recently happed at my work. One of our older machines (physical machine, not software) broke down recently. The company no longer supported or could supply parts for the machine; I believe it was over 20 years old. Through a contact at the company, we got in touch with a semi-retired ex-employee of the company who would repair the machine for us. Apparently he had accumlated many of these machines through auctions etc. to use for parts in his private repair business. The cost wasn't cheap, but it was much cheaper than buying a new machine from the original company, and we were very pleased with the repair service.
However, a third party providing such support is all but impossible with software. A retiree may be more than willing to support software he once worked on on a part time basis, but I imagine that would be all but impossible due to the lack of source code. (Or perhaps it'd just be impractical due to the complexity of software?)
I seem to have failed to fully convey my points. Being the patient type, I'll
explain further.
"What if Excel didn't implement its own window manager and actually allowed one
to view two windows side by side in the fashion one has already learned?"
What if Users could find the "Window|Compare Side by Side" command?
Multiple Excel documents each have an entry in the Windows task bar much
like every other window. Yet they do not behave like other windows. They are
all trapped within the same Excel window. The "X" in the Excel window (not the
smaller document specific "x") closes the Excel window which also closes all
Excel documents within that windows (yes, a close to worthless confirmation
dialog pops up). In my opinion this is very confusing. Notably, another
Office component, Word, behaves in the "normal" way. Perhaps you don't feel
this is an issue, but to it is non-standard. And because it's
non-standard for no apparent reason (is it cruft from a time before
window managers? I don't know, and it doesn't excuse the behavior) it is wrong
(in my opinion of course).
"What if Excel allowed one to save to a folder with a "[" in the name, which
Windows happily allows one to create?"
You know that square brackets have a special use in Excel, right?
Brackets have special meaning in the filename? That's news to me;
please explain. And it's not even the filename. The folder. The folder in
which the Excel file is saved. If one tries to save an Excel file (using Excel
2003 SP1 on WinXP Pro 2002 SP2) in a folder containing brackets, it fails and
goes to the trouble of popping up a handy error message: "The file could not be
accessed. Try one of the following: 1)... 2)... 3) Make sure the file name
does not contain... < > ? [ ] : | or * 4)..." conveniently failing to point
out which of the four options is the culprit and failing to point out
that "file name" also means "folder name".
"What if Excel had a dynamic transpose function?"
There is Edit|Paste Special|Transpose. I can only guess you were looking for
more.
Yes, looking for more... like, maybe I add more data. It won't be
duplicated in the transposed part. Ok, fine; Excel doesn't work that way. But
even worse, the transposed part isn't linked to the original data. Not fine.
I'm sure this one can be overcome with a bit of work, but why does the built-in
transpose refuse to link data?
"What if Excel had used MEAN() instead of AVERAGE()?"
Then there'd be no need for overly picky users to write their own MEAN()
function in order to save three keystrokes; what fun would that be? (You know
you can write your own functions, right?)
Ok, this one's very minor. But I was taught that "average" was a generic
term, most often meaning "mean". The function definition of AVERAGE says
"... arithmetic mean...".
What if Excel didn't implement its own window manager and actually allowed
one to view two windows side by side in the fashion one has already learned?
What if Excel allowed one to save to a folder with a "[" in the name, which
Windows happily allows one to create? What if Excel didn't have math errors
(or so the Gnumeric people claim). What if Excel had a dynamic transpose
function? What if Excel had used MEAN() instead of AVERAGE()?
If GPL3 does what your parent claims it does, then it's entirely in keeping with the spirit of the GPL. The GPL's goal is to require that a binary always include source code (if the use of the binary asks). Requiring the user to violate the DMCA in order to get the source code is clearly a problem.
Ah, I had assumed that the "click agree" was a fault of inflexible Windows
installers, and not really of the actual software authors...
I'm not sure I agree about needing to see a "license to use" however.
Books, lamps, garden tools, etc. do not include a license to use... one may
implicitly do as one pleases (within the law of course) with one's own
property. It seems like the fact that most consumer software requires a
(legally dubious?) license to use, leading free software to also display a
license, is letting the proprietary folks define the terms of debate. No one
complains if "that's just the way things are". And I'd really hate to see this
practice spread to books and garden tools. (there's a guy posting on slashdot
with a sig: "attorney Don Shelky says clickwrap EULAs have been tested and
found valid in most jurisdictions". Is this true? I want it to be like: "want
a contract? get a signature. end of story." Or can a book or CD include a
license requiring one to give up the right to, say, resell? or criticize?.)
I guess one could make the argument that since technically one or more
copies must be made in order to execute the software on a computer then one
does in fact need a license to use... but that seems a bit silly and would
almost certainly fall under fair use and thus not be copyright
infringement.
I'm curious: is there any consumer level proprietary software that does not
have a "license agreement", relying on copyright alone? I think I may buy it
just on principle.
I'm sure you've heard this before... You can take "freedom" to always mean
"your personal freedom". Or you can take it to mean "freedom of society", or a
tangled interconnected web of freedom. I am not free to swing my fist into
your nose, as the saying goes. Is this a bad thing? Or maybe I am free to
punch you, but I then must suffer the consequences of either you punching me or
society kicking me out.
Anyway, I don't think "open source" is about freedom at all. Perhaps you
are talking about the GPL vs. BSD license debate? From everything I've heard,
Stallman is right: open source is based on the "many eyes make bugs shallow"
argument, or "many eyes lead to quicker improvements and better software" (I
thought The Mythical Man-Month disputed this very assertion?). The
executive in Stallman's anecdote was certainly adding more eyes.
Remember that by using hte [sic] software, you are agreeing to a
license of some kind (GPL, Apache...whatever).
Incorrect. For most proprietary software, yes, the license attempts to
govern use and must be "agreed to" prior to using the software (whether this is
legally valid of not I don't know but remain extremely skeptical). Most free
and open source software does not include any license governing use (though it
does include a disclaimer of warranty). The GPL merely stipulates conditions
under which actions that would otherwise be copyright infringement may be
performed. And I don't see how any court could decide that a text edited using
a particular program is then a derivitave work of that program; please correct
me if I am wrong.
I've noticed much free software ported to Windows requires, during
installation, that one click "agree" to the GPL. This annoys me to no end
because I need not agree to the GPL in order to use the software. Perhaps this
common practice has confused you.
So the question is: will this change with digital distribution? I'm
assuming at some point the cost of distributing enormous files with film-like
resolution will approach zero, or at least be much cheaper (and reusable)
compared to film reels.
Greyed out menus with no hint of why they are greyed out are especially annoying... Also, what is the computer supposed to do when it determines you are angry? Suddenly stop sucking?
Another thing that really gets me is keyboard shortcuts. Say, firefox in linux: I repeatedly find myself typing Ctrl-K to delete to the end of the address bar string only to find myself performing a google search. I've almost come to the conclusion that applications should not have interfaces. They should merely plug into "the system" which would completely define the interface which would then have enforced consistency (and the same apps could run within different interfaces, for example one optimized for the blind). Of course, there are big problems with this... like, what to do if "the system" doesn't know how to display what your app needs it to display. The Unix commandline is kinda sorta close to this, but can't handle anything interactive, lacks progress information, works best with text, has limited display abilities, etc.
I'm not sure I understand what keyboard interaction has to do with maximized windows... I assume you're using some keyboard command to switch among windows? Does this not work if the windows are not maximized?
Anecdote: I find that maximized windows slow me down considerably for the simple fact that I can only see the contents of one window at once... To say "a lot of smart geeks do it" is an argument from authority, and not very helpful.
On a slightly different note, I've found that poorly designed webpages combined with resolution dependent windowing systems leads to webpages that just look horrible in a windows that are the "wrong" size, whether the window is maximized and too large, or not maximized and too small. Furthermore, I have a widescreen monitor, and it's a chore to read text from a webpage (or any program for that matter) that spans the entire window due to a maximized window... shifting my eyes great distances becomes tedious. Newspapers use multiple columns for a reason. Books purely of text generally aren't 30" wide either. And no, designing webpages that restrict column widths to 600 pixels isn't a good solution. If it is, what should one do with all the blank space on the sides?
Perhaps tiling window managers are the answer (do you consider the windows to be "maximized" in the case where say, two are tiled side by side? I do not; are we simply disagreeing on terminology?). Sadly many apps [that I use] do not play well with tiling window managers. Nor do I particularly enjoy learning yet another set of keyboard shortcuts and hoping they don't interfere with whatever keyboard shortcuts my app wants... I guess the solution here is better designed apps and window manager.
Question: are they not identical due to external circumstances causing them to grow differently, or due to slight mutations of the DNA, or both? Thanks.
Based on suggestions I've read in the comments, how about this:
Voter enters polling place, name scratched off list as usual. Voter enters
booth. For each office up for election, voter types* a name or names+ into the
voting machine. A blank vote or "Nobody" would indicate no vote for that
office. Referendums etc. could be indicated with some predefined response
(preferably more than a simple "yes" or "no" in order to avoid
Windows-dialog-box-style confusion). When finished, the voting machine prints
out the completed ballot. The format is importantly both human readable and
machine readable via OCR. Surely if the machine knows the font beforehand, OCR
can be fairly quick and highly accurate...? A ballot would essentialy be a list
like:
President: John P. Doe
Senator: Jane T. Smith
Representative: Joe G. Johnson
Increase taxes for schools?: Do not increase
A ballot may contain special marks to help a machine reader align the text,
but the actual vote info must be human readable (i.e. not a barcode). The
voter reviews the ballot and either destroys it and creates a new one, or
submits it to the ballot box. Ballots are then machine tallied after all
ballots are collected (it is important to not tally instaneously for the sake
of voter anonymity). Hand recounts may be conducted as necesarry.
The good parts about this are 1) machine countable, 2) human countable, 3)
transparent (voter puts physical paper ballot into box rather than bits into a
database), 4) tamper resistant (difficult to invalidate votes by marking or
tampering with the ballot after the fact) 5) anonymous.
One problem is: how to type a candidate's name. Keyboard? What about those
with disabilities? I'm not really familiar with alternate text entry systems,
but surely some exist.
* The biggest problem is, of course, determining who is meant by "John P.
Doe", since there may be many John P. Does in America. I don't really like the
idea of requiring people to "get on the ballot" because anyone who doesn't know
who to vote for will almost certainly pick a candidate who is on the ballot.
But I don't really have a solution for an all-write-in system. Please address
this as a separate issue. In lieu of requiring a typed name, the system could
easily offer a selection of candidates as is common now. (How do write-in votes
work now? I assume they are silently ignored unless it's clear that a majority
of votes are not for someone on the ballot which almost surely never
happens).
+ Some offices may allow multiple candidates. Some voting systems may allow
multiple votes, possibly ranked, for a single final winner. This voting method
lends itself well to these alternative (surperior IMO) methods.
Slashdot Mirror
Yeah right, who's the only one here who knows illegal ninja moves from the government? Uppercuts of Mass Destruction.
16 bit typically refers to 16 bits per what I'll call a sample. 24 bit RGB has 8 bits per sample, 3 samples per pixel. 16 bit samples would be 16x3 = 48 bit RGB.
You'd have to design it with "instant-retaliation". Every time you mark a driver, the driver's score goes up one point, but your score also goes up one point. This would discourage people from blindly marking everyone they see. Only cars that had been marked [hopefully] independently would exceed the threshold for a ticket. This system would still be vulnerable to a coordinated attack where a band of trolls each mark the same target. This might be mitigated somewhat by changing the 1-1 retaliatino ratio.
You could also have a meta moderation that was simply based on the rate at which you doled out marks. If you exceed a certain rate (say, two standard deviations above average), then your marks are ignored.
Emacs may well lack some features... but for anti-aliased text one can always run it from an appropriate X terminal emulator. I haven't used it, but there is a version that uses Gtk2 and presumably has anti-aliased text.
Agreed. I think the lawyer's job is not to argue to get "bad guys" off the hook (and perhaps it isn't; I don't see lawyers outside of TV or movies), but to ensure a fair trial.
Given that the Toyota is marginally more fuel efficient than a small gas or diesel car, the only good from buying one seems to be supporting hybrid research with your dollar (which is a good thing). Fuel efficiency folks also neglect the energy required to make the car in the first place. A used (the manufacturing energy is a sunk cost) but fuel efficient car probably wins in total energy. Of course, that's more of a short term thing; not everyone can buy used cars forever. I have no idea how short "short term" means in this case. Newer cars also tend to pollute less due to better catalytic converters, but a two to five year old car will probably be about the same in that department.
In conclusion, I applaud your support of hybrid research. Ride a bike if you can. It's a shame our (U.S.) culture seems so derisive of doing that, and that most cities (at least the ones I've seen) are designed exclusively for cars.
The purpose of the GPL is to ensure that a software program includes the four freedoms. If a program will only function with, say, a signed binary module, and it is impossible for the end use to create such a signed module of his own, then he is missing the "freedom to study how the program works, and adapt it to [his] needs". I don't really see how adding a new restriction to the GPL goes against this spirit of the GPL. It may go against your idea of what a license should be, but to say GPLv3 is a political statement while GPLv2 is not is wrong because the GPLv3 additions are within the spirit of the four freedoms.
Can I sell used digital tracks? Can I lend digital tracks to others deleting my copy while it's lent out? Consider a "hot-potato" or collaborative library style p2p system where each member contributes the songs he owns to the pool. When another member wishes to hear a song, she determines which member or members own the song, borrows (or buys for free) the song (which the owner then deletes), listens to the song, and then either returns it to the owner or registers with the pool as the new owner ready to lend to others.
I wonder what a court would think of this... One would most likely need to come up with a file transfer protocol that could "move" a file rather than "copy" a file (like, transfer bytes 1 to X, wait for confirmation, delete bytes 1 to X, transfer bytes X to 2*X, ...).
Can Intel not license the same Z-RAM patents?
Your comment made me think of a situation that recently happed at my work. One of our older machines (physical machine, not software) broke down recently. The company no longer supported or could supply parts for the machine; I believe it was over 20 years old. Through a contact at the company, we got in touch with a semi-retired ex-employee of the company who would repair the machine for us. Apparently he had accumlated many of these machines through auctions etc. to use for parts in his private repair business. The cost wasn't cheap, but it was much cheaper than buying a new machine from the original company, and we were very pleased with the repair service.
However, a third party providing such support is all but impossible with software. A retiree may be more than willing to support software he once worked on on a part time basis, but I imagine that would be all but impossible due to the lack of source code. (Or perhaps it'd just be impractical due to the complexity of software?)
I seem to have failed to fully convey my points. Being the patient type, I'll explain further.
Multiple Excel documents each have an entry in the Windows task bar much like every other window. Yet they do not behave like other windows. They are all trapped within the same Excel window. The "X" in the Excel window (not the smaller document specific "x") closes the Excel window which also closes all Excel documents within that windows (yes, a close to worthless confirmation dialog pops up). In my opinion this is very confusing. Notably, another Office component, Word, behaves in the "normal" way. Perhaps you don't feel this is an issue, but to it is non-standard. And because it's non-standard for no apparent reason (is it cruft from a time before window managers? I don't know, and it doesn't excuse the behavior) it is wrong (in my opinion of course).
Brackets have special meaning in the filename? That's news to me; please explain. And it's not even the filename. The folder. The folder in which the Excel file is saved. If one tries to save an Excel file (using Excel 2003 SP1 on WinXP Pro 2002 SP2) in a folder containing brackets, it fails and goes to the trouble of popping up a handy error message: "The file could not be accessed. Try one of the following: 1) ... 2) ... 3) Make sure the file name
does not contain ... < > ? [ ] : | or * 4) ..." conveniently failing to point
out which of the four options is the culprit and failing to point out
that "file name" also means "folder name".
Yes, looking for more... like, maybe I add more data. It won't be duplicated in the transposed part. Ok, fine; Excel doesn't work that way. But even worse, the transposed part isn't linked to the original data. Not fine. I'm sure this one can be overcome with a bit of work, but why does the built-in transpose refuse to link data?
Ok, this one's very minor. But I was taught that "average" was a generic term, most often meaning "mean". The function definition of AVERAGE says "... arithmetic mean ...".
What if Excel didn't implement its own window manager and actually allowed one to view two windows side by side in the fashion one has already learned? What if Excel allowed one to save to a folder with a "[" in the name, which Windows happily allows one to create? What if Excel didn't have math errors (or so the Gnumeric people claim). What if Excel had a dynamic transpose function? What if Excel had used MEAN() instead of AVERAGE()?
If GPL3 does what your parent claims it does, then it's entirely in keeping with the spirit of the GPL. The GPL's goal is to require that a binary always include source code (if the use of the binary asks). Requiring the user to violate the DMCA in order to get the source code is clearly a problem.
Indeed. /me wonders how this marketing is differnet from Intel's marketing.
Thanks for clearing that up. For a moment there I thought I'd gone mad...
Ah, I had assumed that the "click agree" was a fault of inflexible Windows installers, and not really of the actual software authors...
I'm not sure I agree about needing to see a "license to use" however. Books, lamps, garden tools, etc. do not include a license to use... one may implicitly do as one pleases (within the law of course) with one's own property. It seems like the fact that most consumer software requires a (legally dubious?) license to use, leading free software to also display a license, is letting the proprietary folks define the terms of debate. No one complains if "that's just the way things are". And I'd really hate to see this practice spread to books and garden tools. (there's a guy posting on slashdot with a sig: "attorney Don Shelky says clickwrap EULAs have been tested and found valid in most jurisdictions". Is this true? I want it to be like: "want a contract? get a signature. end of story." Or can a book or CD include a license requiring one to give up the right to, say, resell? or criticize?.)
I guess one could make the argument that since technically one or more copies must be made in order to execute the software on a computer then one does in fact need a license to use... but that seems a bit silly and would almost certainly fall under fair use and thus not be copyright infringement.
I'm curious: is there any consumer level proprietary software that does not have a "license agreement", relying on copyright alone? I think I may buy it just on principle.
I'm sure you've heard this before... You can take "freedom" to always mean "your personal freedom". Or you can take it to mean "freedom of society", or a tangled interconnected web of freedom. I am not free to swing my fist into your nose, as the saying goes. Is this a bad thing? Or maybe I am free to punch you, but I then must suffer the consequences of either you punching me or society kicking me out.
Anyway, I don't think "open source" is about freedom at all. Perhaps you are talking about the GPL vs. BSD license debate? From everything I've heard, Stallman is right: open source is based on the "many eyes make bugs shallow" argument, or "many eyes lead to quicker improvements and better software" (I thought The Mythical Man-Month disputed this very assertion?). The executive in Stallman's anecdote was certainly adding more eyes.
Incorrect. For most proprietary software, yes, the license attempts to govern use and must be "agreed to" prior to using the software (whether this is legally valid of not I don't know but remain extremely skeptical). Most free and open source software does not include any license governing use (though it does include a disclaimer of warranty). The GPL merely stipulates conditions under which actions that would otherwise be copyright infringement may be performed. And I don't see how any court could decide that a text edited using a particular program is then a derivitave work of that program; please correct me if I am wrong.
I've noticed much free software ported to Windows requires, during installation, that one click "agree" to the GPL. This annoys me to no end because I need not agree to the GPL in order to use the software. Perhaps this common practice has confused you.
Hush PC is somewhat pretty. But it is quite expensive, and I don't think they sell the case alone.
So the question is: will this change with digital distribution? I'm assuming at some point the cost of distributing enormous files with film-like resolution will approach zero, or at least be much cheaper (and reusable) compared to film reels.
Greyed out menus with no hint of why they are greyed out are especially annoying... Also, what is the computer supposed to do when it determines you are angry? Suddenly stop sucking?
Another thing that really gets me is keyboard shortcuts. Say, firefox in linux: I repeatedly find myself typing Ctrl-K to delete to the end of the address bar string only to find myself performing a google search. I've almost come to the conclusion that applications should not have interfaces. They should merely plug into "the system" which would completely define the interface which would then have enforced consistency (and the same apps could run within different interfaces, for example one optimized for the blind). Of course, there are big problems with this... like, what to do if "the system" doesn't know how to display what your app needs it to display. The Unix commandline is kinda sorta close to this, but can't handle anything interactive, lacks progress information, works best with text, has limited display abilities, etc.
I'm not sure I understand what keyboard interaction has to do with maximized windows... I assume you're using some keyboard command to switch among windows? Does this not work if the windows are not maximized?
Anecdote: I find that maximized windows slow me down considerably for the simple fact that I can only see the contents of one window at once... To say "a lot of smart geeks do it" is an argument from authority, and not very helpful.
On a slightly different note, I've found that poorly designed webpages combined with resolution dependent windowing systems leads to webpages that just look horrible in a windows that are the "wrong" size, whether the window is maximized and too large, or not maximized and too small. Furthermore, I have a widescreen monitor, and it's a chore to read text from a webpage (or any program for that matter) that spans the entire window due to a maximized window... shifting my eyes great distances becomes tedious. Newspapers use multiple columns for a reason. Books purely of text generally aren't 30" wide either. And no, designing webpages that restrict column widths to 600 pixels isn't a good solution. If it is, what should one do with all the blank space on the sides?
Perhaps tiling window managers are the answer (do you consider the windows to be "maximized" in the case where say, two are tiled side by side? I do not; are we simply disagreeing on terminology?). Sadly many apps [that I use] do not play well with tiling window managers. Nor do I particularly enjoy learning yet another set of keyboard shortcuts and hoping they don't interfere with whatever keyboard shortcuts my app wants... I guess the solution here is better designed apps and window manager.
I haven't used Reiser much, but I thought this "tail packing" or whatever they call it was optional?
Question: are they not identical due to external circumstances causing them to grow differently, or due to slight mutations of the DNA, or both? Thanks.
Based on suggestions I've read in the comments, how about this:
Voter enters polling place, name scratched off list as usual. Voter enters booth. For each office up for election, voter types* a name or names+ into the voting machine. A blank vote or "Nobody" would indicate no vote for that office. Referendums etc. could be indicated with some predefined response (preferably more than a simple "yes" or "no" in order to avoid Windows-dialog-box-style confusion). When finished, the voting machine prints out the completed ballot. The format is importantly both human readable and machine readable via OCR. Surely if the machine knows the font beforehand, OCR can be fairly quick and highly accurate...? A ballot would essentialy be a list like:
A ballot may contain special marks to help a machine reader align the text, but the actual vote info must be human readable (i.e. not a barcode). The voter reviews the ballot and either destroys it and creates a new one, or submits it to the ballot box. Ballots are then machine tallied after all ballots are collected (it is important to not tally instaneously for the sake of voter anonymity). Hand recounts may be conducted as necesarry.
The good parts about this are 1) machine countable, 2) human countable, 3) transparent (voter puts physical paper ballot into box rather than bits into a database), 4) tamper resistant (difficult to invalidate votes by marking or tampering with the ballot after the fact) 5) anonymous.
One problem is: how to type a candidate's name. Keyboard? What about those with disabilities? I'm not really familiar with alternate text entry systems, but surely some exist.
* The biggest problem is, of course, determining who is meant by "John P. Doe", since there may be many John P. Does in America. I don't really like the idea of requiring people to "get on the ballot" because anyone who doesn't know who to vote for will almost certainly pick a candidate who is on the ballot. But I don't really have a solution for an all-write-in system. Please address this as a separate issue. In lieu of requiring a typed name, the system could easily offer a selection of candidates as is common now. (How do write-in votes work now? I assume they are silently ignored unless it's clear that a majority of votes are not for someone on the ballot which almost surely never happens).
+ Some offices may allow multiple candidates. Some voting systems may allow multiple votes, possibly ranked, for a single final winner. This voting method lends itself well to these alternative (surperior IMO) methods.
Discuss.