ANSI X3.226 in PDF at a bargain
Erik Naggum
ANSI (www.ansi.org) is actually selling a downloadable PDF version of the
Common Lisp standard for only USD 18. The file is fully 65M, so I tend
to suspect that it has been scanned in, especially since C++ is only 2M,
but it should be as close to the real thing as you can virtually get.
Whether it is more or less conveniently readable than HTML is a matter of
personal taste of course, but at least the expense is no longer an excuse
not to have this document available for viewing.
///
Raymond Toy
Erik> While looking for some other standards documents today, I
Erik> discovered that ANSI (www.ansi.org) is actually selling a
Erik> downloadable PDF version of the Common Lisp standard for
Erik> only USD 18. The file is fully 65M, so I tend to suspect
Erik> that it has been scanned in, especially since C++ is only
Erik> 2M, but it should be as close to the real thing as you can
Erik> virtually get. Whether it is more or less conveniently
Erik> readable than HTML is a matter of personal taste of course,
Erik> but at least the expense is no longer an excuse not to have
Erik> this document available for viewing.
Some one mentioned to me a short while back that it is a scan, and a
pretty bad one at that. While I would like a nice paper copy and am
willing to pay a reasonable price, a bad scan for $18 doesn't count.
Ray
Erik Naggum
| Some one mentioned to me a short while back that it is a scan, and a
| pretty bad one at that. While I would like a nice paper copy and am
| willing to pay a reasonable price, a bad scan for $18 doesn't count.
I am sorry to hear that. I have a copy of the paper version (which
self-respecting Common Lisp expert does not? :), and do not really need a
PDF version, but now I want to see this for myself and have started a
download, which does not seem to want to finish. Sigh. In any case, if
this is a poor quality document, perhaps it would be possible to produce
a nicer (and a lot smaller! -- it is reported to be 67,265,589 bytes) PDF
version from the TeX or DVI files. Kent?
///
Kent M Pitman
>
> While looking for some other standards documents today, I discovered that
> ANSI (www.ansi.org) is actually selling a downloadable PDF version of the
> Common Lisp standard for only USD 18. The file is fully 65M, so I tend
> to suspect that it has been scanned in,
Hmm. Well, they were given PostScript files so I can't imagine they didn't
just distill them.
> especially since C++ is only 2M,
Probably just not as many font shifts. It's those font shifts that enabled
all the hyperlinks to be created from the TeX.
> but it should be as close to the real thing as you can virtually get.
> Whether it is more or less conveniently readable than HTML is a matter of
> personal taste of course, but at least the expense is no longer an excuse
> not to have this document available for viewing.
I wonder if the price fell due to its age or due to a change in ANSI policy.
It would be great if ANSI didn't see paper publishing as its source of funding.
Kaz Kylheku
>I wonder if the price fell due to its age or due to a change in ANSI policy.
>It would be great if ANSI didn't see paper publishing as its source of funding.
It's just the policy. Newer documents like C++ 1998 are also available for
eighteen bucks from ANSI's website.
Erik Naggum
| Hmm. Well, they were given PostScript files so I can't imagine they didn't
| just distill them.
The PDF document is a scan. It is _pathetic_, as Raymond Toy said, and I
should have saved my dollars and not wasted the bandwidth. Bummer. I
have not transferred the 50M Ada standard, but I suspect it is just as
bad, and will complain to ANSI for this low-quality product.
| Probably just not as many font shifts. It's those font shifts that
| enabled all the hyperlinks to be created from the TeX.
Nah, the C++ standard is a regular PDF file with real text.
| I wonder if the price fell due to its age or due to a change in ANSI
| policy. It would be great if ANSI didn't see paper publishing as its
| source of funding.
It appears that all electronic standards from ANSI are USD 18, now, but I
have only checked computer-related ones. _I_ think this is fantastic,
but the ISO versions are ten to twenty times more expensive, so at least
one of the behemoths have seen the light.
///
Simon András
> The PDF document is a scan. It is _pathetic_, as Raymond Toy said, and I
> should have saved my dollars and not wasted the bandwidth. Bummer. I
> have not transferred the 50M Ada standard, but I suspect it is just as
> bad, and will complain to ANSI for this low-quality product.
Is version 15.17R very far from the real thing?
Andras
James A. Crippen
> * Kent M Pitman
> | Hmm. Well, they were given PostScript files so I can't imagine they didn't
> | just distill them.
>
> The PDF document is a scan. It is _pathetic_, as Raymond Toy said, and I
> should have saved my dollars and not wasted the bandwidth. Bummer. I
> have not transferred the 50M Ada standard, but I suspect it is just as
> bad, and will complain to ANSI for this low-quality product.
Please do. You're not the only one. I was the guy who told Raymond
Toy that the scans sucked.
I'm fairly certain that ANSI must have lost the original PostScript
doc and just cut one of their paper versions and scanned it with a
sheet feeder. The horrible thing is that they seem to have scanned it
at <= 300 dpi.
As Erik noted, would it be a good idea to take the final TeX version
and generate fresh postscript from it?
'james
--
James A. Crippen <ja...@unlambda.com> ,-./-. Anchorage, Alaska,
Lambda Unlimited: Recursion 'R' Us | |/ | USA, 61.2069 N, 149.766 W,
Y = \f.(\x.f(xx)) (\x.f(xx)) | |\ | Earth, Sol System,
Y(F) = F(Y(F)) \_,-_/ Milky Way.
Erik Naggum
| Is version 15.17R very far from the real thing?
No, apparently not. It seems only some minor administrative glitches
held up the standard's publication for a long time and that nothing
happened editorially in this interval.
///
James A. Crippen
Integrating the post-standard notes in the HyperSpec would be a Good
Thing.
Kent M Pitman
> As Erik noted, would it be a good idea to take the final TeX version
> and generate fresh postscript from it?
IMO it would just be putting money in ANSI's pocket. I'd personally
rather not do that.
The Lisp Community (through it's vendors) paid just shy of a half million
dollars to produce the CL standard, at which point ANSI slapped its
copyright notice on the final result and started charging some outrageous
amount (I can't remember, but I thought it was in the hundreds of dollars).
The people who paid to PRODUCE the document (the vendors) got none of that.
I don't personally see doing ANSI any further financial favors.
If someone wants to make a PDF version available, they should do it from
the final draft. It is identical in all technical respects to the final
version. That's the draft that Franz's CL documentation worked from, for
example. It differs only in headers, edit history, cover page, and "book
design". ("book design" = I saved about 200 pages of dead tree per copy by
fudging the font size and interline whitespace a tiny bit. Sigh. But for
PDF, that's of much less consequence than for ANSI's printed document.)
Incidentally, I don't think I gave ANSI the TeX sources. I gave them DVI
and PostScript, I believe. They should have been able to distill the
postscript to PDF, I'd think; but I'm not surprised they can't generate
a proper copy from sources they don't have.
I have record somewhere of exactly what I gave them. Attempting to
act in the best interest of the Lisp community, my recollection is
that I keep what I gave them to what was actually required. I didn't
want them having sources both so they didn't edit anything, and also
so they didn't start to claim they had ownership of the source by
virtue of having possession of it, since they once over the phone
suggested they might do something vaguely like that. That remark
triggered a multi-month delay while I did things that I felt secured
us a better position. There's a lot of ugly unpublished backstory
here, actually. I have generally declined to go into details about
this, but let me just say it leaves me with no feelings of goodwill
toward ANSI.
Kent M Pitman
It differs only in cover pages, page headers, font size, whitespace.
There is no change in any technically binding part of the standard
between this draft and the actual standard.
Kent M Pitman
> Erik Naggum <er...@naggum.net> writes:
>
> > * Simon András
> > | Is version 15.17R very far from the real thing?
> >
> > No, apparently not. It seems only some minor administrative glitches
> > held up the standard's publication for a long time and that nothing
> > happened editorially in this interval.
>
> Integrating the post-standard notes in the HyperSpec would be a Good
> Thing.
Post-standard notes? There is much mail that has gone back and forth,
but technically I believe there are no formal documents of any kind
that have happened subsequent to publication. Then again, I'm no longer
a J13 member so I can't say with complete certainty what's happened in
the last year or so.
Edward Jason Riedy
-
- I have record somewhere of exactly what I gave them. Attempting to
- act in the best interest of the Lisp community, my recollection is
- that I keep what I gave them to what was actually required. I didn't
- want them having sources both so they didn't edit anything, [...]
While your other comments are dead-on, this one can lead
to nasty problems. The IEEE has "lost" the original nroff
source for 754-1985 and 854-1987. (Unfortunately, so have
the authors and UC Berkeley.) However, this has not stopped
them from recoding it into some other format. In the
process, they have introduced _serious_ errors and typos.
And they failed to fix the ones that were there.
They'll still edit it however they want. Not providing
the source simply induces extra errors when they do so.
Of course, they'd likely scan it into Word or FrameMaker
rather than starting from a "primitive" system like TeX.
yech.
Jason, who had no idea how much politics went into even
exceedingly mathematical standards...
--
Kent M Pitman
> And Kent M Pitman writes:
> -
> - I have record somewhere of exactly what I gave them. Attempting to
> - act in the best interest of the Lisp community, my recollection is
> - that I keep what I gave them to what was actually required. I didn't
> - want them having sources both so they didn't edit anything, [...]
>
> While your other comments are dead-on, this one can lead
> to nasty problems. The IEEE has "lost" the original nroff
> source for 754-1985 and 854-1987.
The files transferred are redundantly stored with several different agencies.
Kim Barrett, X3J13 secretary at the time, has a copy. I have a copy in a
backup tape in a safe deposit box and on my local machine. And anyway,
the technical content (otehr than the frontmatter) is publicly accessible
as the final draft that J13 voted on. And I believe copies also exist in
the Harlequin backup tapes if not also still their live source archives,
since Harlequin was given legal license by ANSI to produce derivative works.
> (Unfortunately, so have
> the authors and UC Berkeley.) However, this has not stopped
> them from recoding it into some other format. In the
> process, they have introduced _serious_ errors and typos.
> And they failed to fix the ones that were there.
Nothing obliges us as a community to accept what they do. And we do have
means of verifying correctness of what they do if it comes down to that.
> They'll still edit it however they want. Not providing
> the source simply induces extra errors when they do so.
I don't agree. I got them to stop before. I claim they don't have
the right to make such changes on their own. I got them to stop cold
in their tracks the last time I asserted that. The issue did come up
during publication.
> Of course, they'd likely scan it into Word or FrameMaker
> rather than starting from a "primitive" system like TeX.
> yech.
I don't really think so. It's possible to do technically. But they
don't have the right. ANSI showed up at our first meeting of J13 and
told us that their first goal was "not getting sued" and their second
goal was "producing standards".
> Jason, who had no idea how much politics went into even
> exceedingly mathematical standards...
And I suspect you still don't. I've not gone into the full detail of it
all. It was really painful and ugly. And my impression from talking
around is that every other language committee has gone through similar
pains. Pity.
Bijan Parsia
[snip]
> I wonder if the price fell due to its age or due to a change in ANSI
> policy. It would be great if ANSI didn't see paper publishing as its
> source of funding.
A data point: The ANSI Smalltalk spec was available as a PDF for $20 since
it was released a few years back, but I think the paper spec was in the
"normal" (i.e., rather high) price range. So, they seem to have generally
moved to selling the official PDF at a vastly cheaper rate.
(I'll note too that the last draft is still available for free. While I'd
really love to have the ANSI spec hard copies for $30-60 at my local
bookstore, it's hard to be *too* whiny about the moves they've made to
expand access.)
Cheers,
Bijan Parsia.
James A. Crippen
> (I'll note too that the last draft is still available for free. While I'd
> really love to have the ANSI spec hard copies for $30-60 at my local
> bookstore, it's hard to be *too* whiny about the moves they've made to
> expand access.)
I don't think that they realize they could make a good chunk of money
if they started selling the specifications of various things (not just
programming languages, since ANSI standardizes a lot of things) in
bookstores. Not all of their standards are 5000 page monsters. A
number of them are certainly bindable into single books. Selling
something like the Lisp spec (which is what, 1500 pages or something?)
for $100 bucks would be great. I'd certainly pay $100 of $150 for the
full Lisp standard (including semistandard stuff like Gray streams,
etc) in hardcopy with a nice hardbound cover and durable binding. I'd
even consider paying $60 or $70 for a not-so-hot binding like that
used for the CLtL2 or those used for those 'SAMS' and 'QUE' computer
books for newbies that include CD-ROMs so they have an excuse to sell
them at higher prices.
Kent M Pitman
> Bijan Parsia <bpa...@email.unc.edu> writes:
>
> > (I'll note too that the last draft is still available for free. While I'd
> > really love to have the ANSI spec hard copies for $30-60 at my local
> > bookstore, it's hard to be *too* whiny about the moves they've made to
> > expand access.)
>
> I don't think that they realize they could make a good chunk of money
> if they started selling the specifications of various things (not just
> programming languages, since ANSI standardizes a lot of things) in
> bookstores. Not all of their standards are 5000 page monsters. A
> number of them are certainly bindable into single books.
To be honest, I don't think our standards are typical. Most standards
are not like programming languages, or traditionally have been. Maybe
it's changing. I don't have access to stats, but I was told by
someone that most are thigns more like 6 page documents which tell you
how far apart the threads on a bolt or nut are supposed to be, and
they're of interest only to vendors, not users, who simply buy the
bols and nuts and expect them to mesh. I think it's anomalous that
programming languages have so many users, and lucrative though it might
seem to those of us who've never published a book to get that extra tens
of thousands of dollars, it may not be worth perturbing ANSI's overall
business structure just to pick up that small amount of money when such
restructuring of the business might cost as much as they stand to make.
I'm just speculating here, and the right answer would only be knowable
to someone with access to budgets and information about the nature of
stnadards. But at least I wanted to point out that the issue may not
be as clearcut as it seems.
> Selling
> something like the Lisp spec (which is what, 1500 pages or something?)
I believe I got the ANSI standard down to just under 1200, though the
draft was about 1400.
> for $100 bucks would be great.
Yes, that's true. At a nickel a page, it'd cost you $45 to print it
yourself, not to mention the cost of your time. (Except Larry Masinter
was kind enough to print some of the drafts for the committee at Xerox,
where they had machines that could print the whole document AND bind it
in something like 8 minutes with no user intervention... I think we had
to replace paper once or twice in the middle, but it was still awesome
to watch that kind of speed...)
> I'd certainly pay $100 of $150 for the
> full Lisp standard (including semistandard stuff like Gray streams,
> etc) in hardcopy with a nice hardbound cover and durable binding.
I'm glad I didn't have to pay. I've used my hardcopy one only a handful
of times. CLHS is basically always faster. It's fun to have a hardcopy,
but... I dunno. I'd rather have the tree back...
cbbr...@acm.org
> I'm glad I didn't have to pay. I've used my hardcopy one only a
> handful of times. CLHS is basically always faster. It's fun to
> have a hardcopy, but... I dunno. I'd rather have the tree back...
You are, of course, slightly biased in the matter :-), as well as,
very likely, having a greater ability to use it than pretty much
anybody!
Hardcopy does have the merit that you can spread it across a table;
video screens lose a _little_ in that regard. And it's much safer to
take a book into the bathroom than a Symbolics machine :-).
But the only reasons I'd want to have a PDF version over the CLHS is
if I intended to:
a) Make sure the CLHS was correct, or
b) Print out the PDF version
Neither of those appear to be useful purposes for having what ANSI is
releasing...
--
(reverse (concatenate 'string "gro.mca@" "enworbbc"))
http://www.ntlug.org/~cbbrowne/sap.html
"Windows: The ``Big O'' of operating systems."
Tim Moore
<cbbr...@acm.org> wrote:
> Kent M Pitman <pit...@world.std.com> writes:
>> I'm glad I didn't have to pay. I've used my hardcopy one only a
>> handful of times. CLHS is basically always faster. It's fun to have a
>> hardcopy, but... I dunno. I'd rather have the tree back...
...
> Hardcopy does have the merit that you can spread it across a table;
> video screens lose a _little_ in that regard. And it's much safer to
> take a book into the bathroom than a Symbolics machine :-).
Notebook computers work for that ;)
Tim
cbbr...@acm.org
> In article <c3Eu7.10980$Uf2.1...@news20.bellglobal.com>, "Unknown"
> <cbbr...@acm.org> wrote:
>
>
> > Kent M Pitman <pit...@world.std.com> writes:
> >> I'm glad I didn't have to pay. I've used my hardcopy one only a
> >> handful of times. CLHS is basically always faster. It's fun to have a
> >> hardcopy, but... I dunno. I'd rather have the tree back...
> > Hardcopy does have the merit that you can spread it across a table;
> > video screens lose a _little_ in that regard. And it's much safer to
> > take a book into the bathroom than a Symbolics machine :-).
>
> Notebook computers work for that ;)
Well, I used ISILO to turn CLTL2 into a Palm document, so there's
another answer...
CLHS was a bit too big, even with the compression scheme, to fit,
unfortunately...
--
(reverse (concatenate 'string "gro.mca@" "enworbbc"))
http://www.cbbrowne.com/info/advocacy.html
Who wants to remember that escape-x-alt-control-left shift-b puts you
into super-edit-debug-compile mode? (Discussion in comp.os.linux.misc
on the intuitiveness of commands, especially Emacs.)
Kent M Pitman
> Kent M Pitman <pit...@world.std.com> writes:
> > I'm glad I didn't have to pay. I've used my hardcopy one only a
> > handful of times. CLHS is basically always faster. It's fun to
> > have a hardcopy, but... I dunno. I'd rather have the tree back...
>
> You are, of course, slightly biased in the matter :-)
Well, IMO, there is no ego involved this. I would not waste my time on the
HyperSpec if paper was faster. And I would not be casually dismissive of
paper if I thought it offered a materially better capability.
Paper is good for a linear read, but I don't expect a huge number of people
to commit such time. Paper is also good when skimming badly index material
and this does *sometimes* happen, but I've always had my paper copy handy
and yet have never reached for it when CLHS was not handy.
The first CLHS you ever saw was version 3 because we tested it in-house
at Harlequin for 1.5 years before releasing it. (It was not for testing
but for internal administrative stupidity that it took it that long to get
out. I was about to cry when I saw CLTL2 come out in webbed form a half
year after I webbed ANSI CL and Harlequin had still not made the decision
to release the product. I thought I'd die an unknown and irrelevant
participant for their failure to move on it in a timely fashion. Had CLHS
come out at the time it was created, it would have been not only the most
comprehensive Lisp document but it would have put Harlequin and Lisp on the
web map because it was surely the largest web document anywhere at the time
it was originally created. Very, very sad. I mention all this now only
because that part of Harlequin which made these decisions is now gone and
replaced by what I regard as substantially more business-aware management,
so you shouldn't hold it against Xanalys that I tell this story.)
In spite of the negative aspects of taking so long to get a release out,
it had a LOT of usage testing before going out and was better for it.
And, to be honest, even after I started using it, I used to sigh when I
thought of looking anything up. I hated the idea of starting a browser
and clicking through. But no matter how I grumbled, gradual experience
told me that it was still faster than reaching for paper.
(Now, I MAY be biased because I know where to look. And a web search facility
would help. But paper has no search facility either, so ... And in the end,
when you don't know where you're going, the ability to click around and
to use Back still outweighs other issues for me.)
"Bottom line" (drat, there are lines below this. well, it's an idiom):
Feel free to accuse me of being anomalous data because I have
special knowledge others don't. But don't think I'm not just as critical,
probably much moreseo, than you are about CLHS.
, as well as,
> very likely, having a greater ability to use it than pretty much
> anybody!
Usage? No. To know where I'm going. Maybe sometimes. Though
sometimes it works against me because I *think* I know and I go the
wrong place instead of going to the obvious place. :-) But probably
that's not the norm.
> Hardcopy does have the merit that you can spread it across a table;
1195 page documents don't spread so well across the table. You have to
break the binding to lie it flat. As often as not, it folds back up on you.
And you can't do BACK.
> video screens lose a _little_ in that regard. And it's much safer to
> take a book into the bathroom than a Symbolics machine :-).
A laptop with CLHS and a running Lisp weighs less to carry into the bathroom
than the spec by itself. ;-)
> But the only reasons I'd want to have a PDF version over the CLHS is
> if I intended to:
> a) Make sure the CLHS was correct, or
Fair enough. I'd never argue with this position.
> b) Print out the PDF version
It would cost you $45 at 5cents a page to print the version available to you.
And think of the impact on the environment. Figure out how many lookups
you'll need this to feel comfortable that those trees died for something.
Really. I had to watch about more than a dozen of those drafts get printed
and distributed ot a dozen or so people, knowing most people would probably
skim a few pages and ignore most of it, perhaps at most using it as a handy
prop to raise the height of a monitor... It was quite sad.
> Neither of those appear to be useful purposes for having what ANSI is
> releasing...
I do agree with this final point, absolutely.
In fact, what's weird is that we talked to ANSI about this at the time
of publication and their view was that they were absolutely not in the
business of providing multiple views on a document. They couldn't go with
HTML because it would reformat things and they couldn't adequately insure
quality. It's not surprising to me that they like PDF since at least it
doesn't do creative layout changing as you adjust the screen, and it
doesn't give you font variability and such. However, that they would PDF
a scanned document is stunning to me from the point of view of what I
understand about their quality concerns. The key issue to them, as they
expressed it, was that there has to be a definitive version and that version
has the name ANSI on it. For them to make two possibly divergent forms
and to put the name ANSI (and the same title page, rather than perhaps a
"master version" and "cheapy not-quite-master version" label) is very bizarre
to me. But I suspect they're reacting desperately to the sudden change in
the way business was being done literally risking putting them out of
business if they didn't do something to adapt. So maybe some ordinary
caution got deferred or something for a time pending finding actual
observed problems.
Eli Barzilay
> Well, I used ISILO to turn CLTL2 into a Palm document, so there's
> another answer...
>
> CLHS was a bit too big, even with the compression scheme, to fit,
> unfortunately...
As a side note, I have a text-only version of CLTL2, I find it way
more convenient to search things -- especially when you search for
usage examples. If anyone is interested, it's available at
http://www.cs.cornell.edu/eli/books/cltl/CLtL-text
[Maybe someday I'll have that amount of time again to do CLHS...]
--
((lambda (x) (x x)) (lambda (x) (x x))) Eli Barzilay:
http://www.barzilay.org/ Maze is Life!
Eli Barzilay
> Well, I used ISILO to turn CLTL2 into a Palm document, so there's
> another answer...
>
> CLHS was a bit too big, even with the compression scheme, to fit,
> unfortunately...
As a side note, I have a text-only version of CLTL2. I find it way
more convenient to search things -- especially when you search for
code examples, or generally any mention of whatever. If anyone is
interested, it's at
Neil Schemenauer
I downloaded the 15.17R files from ftp://parcftp.xerox.com:/pub/cl
and created a PDF file using the following commands:
dviconcat -o cl-spec.dvi <dvi files in order>
dvips -Ppdf -G0 cl-spec.dvi -o cl-spec.ps
ps2pdf -dMaxSubsetPct=100 -dCompatibilityLevel=1.2 \
-dSubsetFonts=true -dEmbedAllFonts=true cl-spec.ps
The resulting PDF file is about 4 MB in size, contains only type
1 (scalable) fonts, and looks good in Acrobat Reader. I think
you need both Ghostscript 6 and the type 1 Computer Modern fonts
from AMS in order to get good results.
Can I make this PDF file available? Alternatively, does someone
know the people running parcftp.xerox.com? Perhaps they could
put a PDF version beside the DVI files.
Neil
James A. Crippen
> Hardcopy does have the merit that you can spread it across a table;
> video screens lose a _little_ in that regard. And it's much safer to
> take a book into the bathroom than a Symbolics machine :-).
Actually, you *can* take a LispM in the bathroom, just make sure you
lift with your legs.
Seriously, you can get one of those el-cheapo NCD X terminals and hook
it up to your network. Boot the thing off your PC or whatever and
then run an X session from the Symbolics (or Explorer, they both
supported this).
Certainly the keyboard isn't as great, but the graphics are probably a
little bit better (or less blurry). And you can read stuff in DocEx
while your legs go numb! :-)
cbbr...@acm.org
>
> > Hardcopy does have the merit that you can spread it across a table;
> > video screens lose a _little_ in that regard. And it's much safer to
> > take a book into the bathroom than a Symbolics machine :-).
> Actually, you *can* take a LispM in the bathroom, just make sure you
> lift with your legs.
> Seriously, you can get one of those el-cheapo NCD X terminals and
> hook it up to your network. Boot the thing off your PC or whatever
> and then run an X session from the Symbolics (or Explorer, they both
> supported this).
> Certainly the keyboard isn't as great, but the graphics are probably
> a little bit better (or less blurry). And you can read stuff in
> DocEx while your legs go numb! :-)
I'm almost sorry I started this thread; hopefully nobody can respond
usefully to the TOUGH question:
What happens if you accidentally drop the mouse?
:-)
--
(reverse (concatenate 'string "gro.mca@" "enworbbc"))
http://www.ntlug.org/~cbbrowne/unix.html
Rules of the Evil Overlord #111. "I will offer oracles the choice of
working exclusively for me or being executed."
<http://www.eviloverlord.com/>
Kent M Pitman
> cbbr...@acm.org writes:
>
> > Well, I used ISILO to turn CLTL2 into a Palm document, so there's
> > another answer...
> >
> > CLHS was a bit too big, even with the compression scheme, to fit,
> > unfortunately...
>
> As a side note, I have a text-only version of CLTL2. I find it way
> more convenient to search things -- especially when you search for
> code examples, or generally any mention of whatever. If anyone is
> interested, it's at
> http://www.cs.cornell.edu/eli/books/cltl/CLtL-text
>
> [Maybe someday I'll have that amount of time again to do CLHS...]
As a total aside, of course, if there had been a way to put search into HTML,
I'd have done it. I was hardline about saying that CLHS had to be an
HTML-only product, not an HTML+HTTP+Special-Server product. It is a crying
shame that HTML doesn't have text search built into it and there's no
reason search capability couldn't have built into the HTTP protocol, which
was designed from the beginning to support what was in HTML.
Marc Spitzer
> ja...@unlambda.com (James A. Crippen) writes:
>> cbbr...@acm.org writes:
>>
>> > Hardcopy does have the merit that you can spread it across a table;
>> > video screens lose a _little_ in that regard. And it's much safer to
>> > take a book into the bathroom than a Symbolics machine :-).
>
>> Actually, you *can* take a LispM in the bathroom, just make sure you
>> lift with your legs.
>
>> Seriously, you can get one of those el-cheapo NCD X terminals and
>> hook it up to your network. Boot the thing off your PC or whatever
>> and then run an X session from the Symbolics (or Explorer, they both
>> supported this).
>
>> Certainly the keyboard isn't as great, but the graphics are probably
>> a little bit better (or less blurry). And you can read stuff in
>> DocEx while your legs go numb! :-)
>
> I'm almost sorry I started this thread; hopefully nobody can respond
> usefully to the TOUGH question:
>
> What happens if you accidentally drop the mouse?
>
>:-)
well it depends where you drop it.
marc
Cor Gest jr
> > Certainly the keyboard isn't as great, but the graphics are probably
> > a little bit better (or less blurry). And you can read stuff in
> > DocEx while your legs go numb! :-)
>
> I'm almost sorry I started this thread; hopefully nobody can respond
> usefully to the TOUGH question:
>
> What happens if you accidentally drop the mouse?
>
Well, it's said often:
Garbage in is Garbage out ....;-)
cor
--
An operatingsystem is just a name you give to the rest of bloating
idiosyncratic machine-based-features you left out of your editor
DMCA, SSSCA, W3C? Who cares? http://thefreeworld.net/ (volunteers needed)
Tim Bradshaw
> Garbage in is Garbage out ....;-)
It's really weird reading this thread in the UK (well, I'm not in the
UK, but my mind is). `Go to the bathroom' means `go to the room where
you have baths, probably to have a bath' here. As someone who often
reads in the bath I've been wondering about all the wrong issues...
--t
Aaron J Reichow
> Erik Naggum <er...@naggum.net> writes:
>
> > * Kent M Pitman
> > | Hmm. Well, they were given PostScript files so I can't imagine they didn't
> > | just distill them.
> >
> > The PDF document is a scan. It is _pathetic_, as Raymond Toy said, and I
> > should have saved my dollars and not wasted the bandwidth. Bummer. I
> > have not transferred the 50M Ada standard, but I suspect it is just as
> > bad, and will complain to ANSI for this low-quality product.
>
> Please do. You're not the only one. I was the guy who told Raymond
> Toy that the scans sucked.
>
> I'm fairly certain that ANSI must have lost the original PostScript
> doc and just cut one of their paper versions and scanned it with a
> sheet feeder. The horrible thing is that they seem to have scanned it
> at <= 300 dpi.
>
> As Erik noted, would it be a good idea to take the final TeX version
> and generate fresh postscript from it?
Well jeeze. With the little doc-making I've done using Acrobat 3, it does
OCR. All they have to do is click a big "convert to text" button, and
Acrobat does it. I suppose they don't want to proof the OCR version, but
isn't that what you're paying for? The OCR was pretty acurate even in the
old version I use at work- you'd think for a printed hardcopy, the OCR
would go pretty slickly.
Regards,
Aaron
Aaron Reichow :: UMD ACM Pres :: http://www.d.umn.edu/~reic0024/
"life, probably the biggest word i've ever said, that says a lot,
because there's a whole lot of words inside my head.." :: atmosphere
Cor Gest jr
Yes indeed, I am always suprised about the size the lu's have in the UK ;-))))
Kent M Pitman
> Well jeeze. With the little doc-making I've done using Acrobat 3, it does
> OCR. All they have to do is click a big "convert to text" button, and
> Acrobat does it. I suppose they don't want to proof the OCR version, but
> isn't that what you're paying for? The OCR was pretty acurate even in the
> old version I use at work- you'd think for a printed hardcopy, the OCR
> would go pretty slickly.
They aren't QUALIFIED to proof it. They are not a body of experts;
they use out-of-house experts. And anyone they might hire to do it
would be aware of the financial stake they had and would charge them a
commensurate fee if they had any sense. By never compensating anyone
fairly for the creation of the standard, they haven't exactly built
any goodwill, and they were stubborn about the production of same at
the time of the original document. I suspect they know they're in a
no-win situation of their own devising and they just decided to
blunder through. I have no sympathy.
ANSI's own stated model of the whole standards process is that we
contracted them, not vice versa. That is, they are more like a lawyer
or stock broker who is done getting paid and looking to churn
business. Now, if they adopted the model that was more appropriate to
them which is that they are like ACM or even Addison-Wesley or Digital
Press and were just a publishing house, then at least it would make
sense that they would call you later about more opportunities since
their business is not standards but publishing, and they'd be willing
to compensate you since it affects their admitted business. I think a
lot of this is caused by them either not having a scientific (if
you'll pardon the meta-use of this overloaded term) understanding of
their business arena or, perhaps worse, their having such an
understanding but not wanting to project that model to customers.
Other examples of businesses that have this mismatch that I file in
the same bucket of my brain's analogy web are Franz (who I think is
selling customer solutions not software, and so surprises people with
their prices that are perfectly in-line for customer solutions but are
high for just software), Disney (which advertises itself as an amusement
park at which you can buy souvenirs but that I'd be surprised if it
didn't know for a certainty internally was a souvenir sales organization
with a gimmick for creating memories that REQUIRED souvenirs), and the
ACM (which advertises itself as a programmer advocacy group but mostly
is just, like ANSI, a publishing house with a gimmick for making people
feel comfortable with giving them valuable free content on which they can
make money without having to pay for said content). [Btw, Franz is unique
among these in that I don't regard their mismatch between business and
presentation as cynically devised, but rather accidental; and, of course,
their placement in this list is purely my opinion. But especially Disney
I don't think is by accident. ACM and ANSI might or might not just be
bumbling...] I don't raise this issue to be critical of any of these
organizations but rather to give some texture to what I think is a business
paradigm of how some companies succeed either by camouflage or accident,
while causing only confusion (sometimes happy, sometimes not) in the part
of their users.
(There's also an intermediate camp where you can see what's
going on and it's more neutral as to which way to go: Kleenex and
Hallmark Cards. The former, if you listen to their ads, sells the
ability to say "Bless You" and hand someone a mark of caring, not a piece
of dead tree to put snot on; the latter sells a sense of family, caring,
memories, comfort, etc. -- not dead trees. But in this category, the issues
are more transparent to the customer who cares...)
Ok, I'm pretty far afield, so should stop. But I hope I've justified why we
should be careful about "helping ANSI" when it may not be apparent what ANSI
really is and what our relationship is to it. ANSI doesn't provide any
services to us that we don't pay through the teeth for. It's utterly
self-service and merely sells use of its name. It also says it's a "voluntary"
standards organization and that we could, if we wanted, pick another. But
that's another gripe I won't bore you with here; suffice it for me to just
allude for the moment to the fact that only ANSI has the special and legally
authorized monopoly of representing the US in international standards, so
it's not obvious where to shop if you don't like "voluntarily" using them.
Marcin Tustin
<cbbr...@acm.org> wrote in message
news:FxRu7.25542$S_6.2...@news20.bellglobal.com...
> I'm almost sorry I started this thread; hopefully nobody can respond
> usefully to the TOUGH question:
>
> What happens if you accidentally drop the mouse?
>
One word for you: marigolds.
Marcin Tustin
Kent M Pitman <pit...@world.std.com> wrote in message
news:sfwwv2b...@world.std.com...
Isn't the IEEE ISO-affiliated (or at least a producer of standards [I am
aware that they have their own LISP]). Also Microsoft (Or was it Netscape?)
got ECMAScript standardised through the European Computer Manufacturers'
Association.
Thomas F. Burdick
> Well jeeze. With the little doc-making I've done using Acrobat 3, it does
> OCR. All they have to do is click a big "convert to text" button, and
> Acrobat does it. I suppose they don't want to proof the OCR version, but
> isn't that what you're paying for? The OCR was pretty acurate even in the
> old version I use at work- you'd think for a printed hardcopy, the OCR
> would go pretty slickly.
Have you ever tried that on something with nice typesetting?
Wordprocessors produce such naïve type that a CS undergrad can write
decent OCR for it. Even a very high quality scan of well-set type, in
my limited experience, is a poor candidate for OCR. I imagine that
very high quality, extremely expensive OCR systems could deal with it,
but the kind of cheapo stuff you get in consumer products probably
wouldn't fare well with technical TeX output. Not to mention that it
sounds like they scanned it < 300 dpi.
Kent M Pitman
> Isn't the IEEE ISO-affiliated (or at least a producer of standards [I am
> aware that they have their own LISP]). Also Microsoft (Or was it Netscape?)
> got ECMAScript standardised through the European Computer Manufacturers'
> Association.
My understanding, and it may not be perfect, is that ANSI contains the
mechanism for this coordination, but that non-ANSI standards can be
blessed by fast-track through ANSI or otherwise coordinated through ANSI.
The member agencies in IEEE are companies. The member agencies in
ANSI are companies. But the member agencies in ISO are countries, and
our country has, as far as I know, only one agency whose job it is to
stand as the representative to ISO. I'm sure they have lots of
"checks and balances" to try to keep things "fair" (by some definition
of fair, as if there could be any such thing as "fair" other than
simply definitionally), but as far as I know, those checks and
balances do not enable me to "voluntarily decline to deal with or
through ANSI" if I want to "voluntarily participate in ISO"; well, to
be "fair", I can't participate in ISO because I'm not a country. So
technically there is no lack of fairness there. Hmm... Well, you see
a hint of my problem with the notion of "fair" and why I talk about it
being merely a definitional issue.
Kent M Pitman
What's funny about this is that in the process of editing the standard, there
was a discussion about how usually you submit a standard to the ANSI editors
and then they reply with myriad editorial requirements. But I found out they
had learned with language standards that they were out of their league and
that they couldn't possibly ever understand the issues of programming syntax
and when it was in play in the running text and when not, and so they finally
decided they would just bless literally whatever the committee did, typos and
all, rather than risk destroying it by asking that anything be changed at all.
(The only thing they were brave enough to ask about was page references in the
index, which they checked thoroughly, turning up bugs in the ad hoc raw TeX
macros we were using where it incorrectly anticipated the page of something
not realizing it would fall to the next page--we ultimately just manually
blundered past that.)
Anyway, having learned that lesson, and if they had any knowledge of the error
rate in OCR, you'd think they'd have applied it to the OCR/PDF exercise and
said "we can't risk this either". But I guess not.
There seem to be just so many reasons for them not to have done this.
I am more and more surprised as I think about this that they proceeded.
Shannon Spires
<pit...@world.std.com> wrote:
> As a total aside, of course, if there had been a way to put search into
> HTML, I'd have done it. I was hardline about saying that CLHS had to be an
> HTML-only product, not an HTML+HTTP+Special-Server product. It is a
> crying shame that HTML doesn't have text search built into it and there's no
> reason search capability couldn't have built into the HTTP protocol,
> which was designed from the beginning to support what was in HTML.
Google does a pretty good job:
http://www.nmia.com/~svspire/lisp/clhs-search.html
This is not what you wanted since it's not self-contained, but I find
it handy.
Shannon Spires
(format nil "svspire~Cnmia.~A" (code-char 64) (reverse "moc"))