Time |
Nickname |
Message |
00:10
π
|
SketchCow |
A magic wand scanner is not a good way to scan magazines, godane. |
00:10
π
|
mistym |
Arg. Someone posted a Laseractive brochure I'd love to buy and scan, but seller ships only to the US. |
00:11
π
|
mistym |
(ebay) |
00:11
π
|
balrog_ |
link? |
00:11
π
|
balrog_ |
pm me ;) |
00:14
π
|
godane |
SketchCow: its the best i can do |
00:15
π
|
SketchCow |
Godane, I will buy you a scanner. |
00:15
π
|
dashcloud |
godane: if you have lots of time, but not much money, you could try the cardboard box scanner here: http://www.instructables.com/id/Bargain-Price-Book-Scanner-From-A-Cardboard-Box/ |
00:15
π
|
SketchCow |
Or he can let me buy him a scanner. |
00:16
π
|
SketchCow |
e-mail jscott@archive.org your shipping address, I will send a scanner to your home |
00:16
π
|
SketchCow |
And I will pay for VueScan, the only good scanning software |
00:16
π
|
dashcloud |
I admit to being mostly ignorant of scanning software, but what puts VueScan ahead of everything else? |
00:17
π
|
balrog_ |
VueScan is better than most manufacturers' scanner software |
00:17
π
|
balrog_ |
gives you more control over the scanner, mainly |
00:17
π
|
SketchCow |
For large scale automatic scanning, it's fantastic. |
00:17
π
|
SketchCow |
No, it also has great automatication |
00:17
π
|
SketchCow |
Scan, BANG, scan, BANG |
00:17
π
|
SketchCow |
I scanned 7000 pages of documents of Infocom with it |
00:18
π
|
godane |
what type of scanner are we talking about? |
00:18
π
|
balrog_ |
one at a time or sheetfed? |
00:18
π
|
SketchCow |
One at a time |
00:18
π
|
balrog_ |
/some/ scanner manufacturers' software is decent and lets you do that. most is archaic and a pain to deal with. |
00:18
π
|
SketchCow |
I am not going to sheetfeed one of a kind memos and writings by Steve Meretzky! |
00:18
π
|
balrog_ |
hey, neither would I! |
00:19
π
|
godane |
i will test out the magic wand scanner first |
00:19
π
|
godane |
then we can talk about some sore of book scanner |
00:19
π
|
balrog_ |
one of my longer term goals is to build a google-style book scanner |
00:21
π
|
godane |
SketchCow: what type of scanner do you have in mind? |
00:22
π
|
godane |
cause some of the book scanners are like alot of money to me |
00:22
π
|
chronomex |
16:14:55 <@SketchCow> Godane, I will buy you a scanner. |
00:22
π
|
chronomex |
$0 |
00:22
π
|
SketchCow |
Yeah, sounds like a discount |
00:24
π
|
dashcloud |
balrog_: if you plan to build a book scanner, go visit diybookscanner.org - tons of models and lots of knowledge there |
00:25
π
|
balrog_ |
dashcloud: this thing http://hackaday.com/2012/11/16/google-books-team-open-sources-their-book-scanner/ |
00:26
π
|
godane |
anyways i just found out i have star wars in japan covered on aots |
00:27
π
|
godane |
luckly to get it since its in a episode from 2008 |
00:27
π
|
godane |
darm it |
00:27
π
|
godane |
there was 30min special of star wars in japan |
02:34
π
|
joepie91 |
SketchCow: huh, what's so awesome about VueScan? |
02:44
π
|
SketchCow |
Well, as I hope you're aware of, most scanner wrap-in software is shit. Shit shit shitty shit shit |
02:44
π
|
SketchCow |
So this is better, on a general sense. |
02:44
π
|
SketchCow |
But more than that, it's really good with large jobs, i.e. you want to scan 150 pages of something. |
02:45
π
|
SketchCow |
It lets you basically be pressing the space bar, scan, puts it in a chronologically sound filename, puts it out in one or two formats, and then done. |
02:45
π
|
SketchCow |
So you can have the softare doing the work of tracking things. |
02:46
π
|
balrog_ |
Epson's wrap-in software is not shit. unfortunately can't say the same about most others. |
02:48
π
|
balrog_ |
I've also used OS X Image Capture which works great if you have a stable version (ancient versions can leak memory, and one version would overwrite rather than chronologically naming files but they fortunately fixed that) |
02:49
π
|
balrog_ |
scanner software has never been great in general |
02:49
π
|
balrog_ |
it seems to be a longstanding trend that is not changing very quickly ... if at all |
02:49
π
|
balrog_ |
pack-in scanner software I should say |
02:52
π
|
godane |
is there any good free software that does this? |
02:53
π
|
balrog_ |
vuescan is very reasonably priced fortunately |
03:12
π
|
godane |
uploaded: http://archive.org/details/The.Screen.Savers.2004.01.16 |
03:32
π
|
SketchCow |
Anyway, yes, some scanner software is good here and there. Vuescan is excellent. |
03:36
π
|
dashcloud |
Do you have a favorite OCR software as well? |
03:40
π
|
SketchCow |
I don't tend to use it, no. |
03:44
π
|
balrog_ |
I've just been dragging things onto OCRKit lately |
03:44
π
|
balrog_ |
but doesn't IA derive do OCR? I'm pretty sure the answer is yes |
03:46
π
|
chronomex |
it does |
03:46
π
|
SketchCow |
It does it really, really poorly |
03:46
π
|
SketchCow |
that shit needs supervision |
03:47
π
|
godane |
agree |
03:47
π
|
godane |
i have seen the ocr on IA be very bad |
03:48
π
|
balrog_ |
I see OCR as mainly an aid for full text search. |
03:48
π
|
balrog_ |
it's just not good enough for making readable text versions :( |
03:50
π
|
chronomex |
yup |
03:58
π
|
dashcloud |
for documents I've needed converted from scanned PDFs back to Word documents, ABBYY's PDF Transformer has done a good job- it's even pretty good about preserving the layout |
03:59
π
|
balrog_ |
but how accurate is the OCR with poor original quality? |
03:59
π
|
balrog_ |
like typewritten stuff |
04:00
π
|
SketchCow |
http://i.imgur.com/5M5Yn.jpg |
04:02
π
|
dashcloud |
I don't know- if you've got some samples, I'll run them through tomorrow for you, and share the results with you |
04:03
π
|
joepie91 |
<SketchCow>It lets you basically be pressing the space bar, scan, puts it in a chronologically sound filename, puts it out in one or two formats, and then done. |
04:03
π
|
dashcloud |
wow- that's impressive |
04:03
π
|
joepie91 |
I actually wrote a script for this |
04:03
π
|
mistym |
balrog_: I've used ABBYY (an older release) w/ typewritten material with good results. It was fairly clean typewritten material mind you |
04:03
π
|
balrog_ |
perhaps http://dec8.info/Digelec/Digelec%20Model%20804%20Manual.pdf |
04:03
π
|
joepie91 |
http://git.cryto.net/cgit/scantools/tree/scan-book |
04:03
π
|
joepie91 |
it's really stupidly simple |
04:04
π
|
joepie91 |
run script, define initial parameters, hit enter for every page |
04:04
π
|
joepie91 |
uses SANE |
04:04
π
|
balrog_ |
mind you I'm connected via mobile from here |
04:04
π
|
balrog_ |
SANE... it would not drive my scanner's feeder properly |
04:04
π
|
DFJustin |
I've used irfanview's batch scan feature in the past as a freebie option |
04:04
π
|
balrog_ |
would do one page and hang ;( |
04:04
π
|
joepie91 |
:/ |
04:04
π
|
joepie91 |
anyway, if anyone has a use for that script |
04:04
π
|
joepie91 |
git clone http://git.cryto.net/repo/projects/joepie91/scantools |
04:05
π
|
mistym |
balrog_: This book had surprisingly clean OCR, for example - http://images.ourontario.ca/brant/1707750/page/6?n= |
04:05
π
|
joepie91 |
feel free to reuse or repurpose in any manner |
04:05
π
|
joepie91 |
unrelated, I just found horrible terrible evil PHP code |
04:05
π
|
joepie91 |
http://php.net/manual/en/function.define.php#109761 |
04:06
π
|
joepie91 |
this guy is *actually* using an ellipsis character in a function name |
04:06
π
|
joepie91 |
(that, and he's assuming a 30 day month instead of using date/time functions which may range, based on purpose, from "not a very good idea" to "oh god you're so fucking stupid and everything you write is going to break") |
04:07
π
|
mistym |
joepie91: ewww |
04:07
π
|
joepie91 |
yes, that was my thought |
04:07
π
|
balrog_ |
ughhhhhhh |
04:08
π
|
balrog_ |
just php makes me shudder |
04:08
π
|
balrog_ |
and bad php? :( |
04:08
π
|
joepie91 |
well, I'm not quite sure what to find more disgusting here |
04:08
π
|
joepie91 |
the fact that guy uses an ellipsis in his function name |
04:09
π
|
joepie91 |
or the fact that PHP allows that crap |
04:09
π
|
joepie91 |
:| |
04:09
π
|
db48x |
heh |
04:09
π
|
db48x |
there are other languages that are generous in what characters you can use in identifiers |
04:09
π
|
db48x |
javascript, for one |
04:10
π
|
joepie91 |
I don't think javascript allows unicode chars |
04:10
π
|
db48x |
lisp is traditionally very generous, but predates unicode |
04:10
π
|
joepie91 |
anyway, javascript is javascript |
04:10
π
|
joepie91 |
you can't really compare it with anything |
04:10
π
|
joepie91 |
:| |
04:10
π
|
db48x |
yes, you can use unicode in identifiers in javascript |
04:10
π
|
joepie91 |
it doesn't work like other languages, also doesn't not work like other languages |
04:10
π
|
balrog_ |
php allows for that crap and is immensely popular |
04:10
π
|
db48x |
yea, I strongly dislike php, but not for this reason |
04:11
π
|
mistym |
We're lucky he didn't decide to use a UTF8 non-breaking space instead. |
04:11
π
|
balrog_ |
I dislike php for its inconsistency and for how annoying it tends to be to work with as opposed to py for example |
04:11
π
|
joepie91 |
db48x, nope.avi: http://owely.com/9dz2kB |
04:11
π
|
balrog_ |
I dislike perl for its unreadability |
04:12
π
|
db48x |
I mean, if I were russian/greek/hindi/chinese/japanese/coptic/whatever and wanted to program, I would probably end up learning english, but I shouldn't be forced to use it |
04:12
π
|
joepie91 |
also, PHP is terrible for inconsistency primarily |
04:12
π
|
joepie91 |
and part of the standard lib is absolute balls |
04:12
π
|
joepie91 |
such as the part that handles date/time |
04:12
π
|
joepie91 |
which is basically impossible to use without abstracting it away, if you wish to preserve your sanity |
04:13
π
|
db48x |
s/and part of// |
04:16
π
|
db48x |
joepie91: oh, and in javascript identifiers have to start with a unicode letter, dollar sign, underscore, or an escaped unicode character |
04:16
π
|
db48x |
see section 7.6 of ECMA-262 |
04:16
π
|
db48x |
:: |
04:16
π
|
db48x |
IdentifierPart |
04:16
π
|
db48x |
IdentifierStart |
04:16
π
|
db48x |
UnicodeCombiningMark |
04:17
π
|
db48x |
UnicodeDigit |
04:17
π
|
db48x |
UnicodeConnectorPunctuation |
04:17
π
|
db48x |
<ZWNJ> |
04:17
π
|
db48x |
<ZWJ> |
04:17
π
|
joepie91 |
db48x: http://owely.com/2LK2D9 |
04:17
π
|
joepie91 |
:P |
04:18
π
|
db48x |
yea, that's not a unicode connector punctuation |
04:19
π
|
joepie91 |
anyway, clearly you can't use this stuff in a function name in JS, haha |
04:19
π
|
joepie91 |
at least not in Chrome |
04:19
π
|
joepie91 |
and Chrome is reference implementation as far as I'm concerned |
04:19
π
|
db48x |
the ellipsis character is in class Po, not class Pc |
04:19
π
|
joepie91 |
seeing as it's the only implementation that is both fast and complete |
04:20
π
|
db48x |
UnicodeConnectorPunctuation :: |
04:20
π
|
db48x |
any character in the Unicode category Γ’ΒΒConnector punctuation (Pc) |
04:20
π
|
DFJustin |
https://kutruff.wordpress.com/2011/02/08/using-%E0%B2%A0_%E0%B2%A0-to-throw-exceptions/ |
04:20
π
|
db48x |
DFJustin: haha, that's awesome |
04:20
π
|
joepie91 |
DFJustin: the URL actually botched my URL bar :( |
04:21
π
|
joepie91 |
http://owely.com/3soHYN |
04:21
π
|
db48x |
you use chrome, what did you expect? |
04:21
π
|
db48x |
I've only used unicode identifiers a few times |
04:21
π
|
DFJustin |
firefox does that too |
04:22
π
|
DFJustin |
it's way more friendly if you're browsing as a japanese user or what have you |
04:22
π
|
joepie91 |
db48x: chrome normally functions very well |
04:22
π
|
db48x |
I used mu in some code that handled microseconds, for instance |
04:22
π
|
chronomex |
nice |
04:22
π
|
joepie91 |
better than a certain other popular open-source browser that I shall not name |
04:22
π
|
db48x |
heh |
04:23
π
|
* |
joepie91 mumbles something about a fox and lack of optimization |
04:23
π
|
joepie91 |
fun: google for ΓΒ’?ΓΒ¦ |
04:24
π
|
joepie91 |
er |
04:24
π
|
joepie91 |
yeah well, ellipsis |
04:24
π
|
db48x |
lol |
04:24
π
|
joepie91 |
forgot that unicode in this client goes derp under WINE |
04:26
π
|
db48x |
what client do you use? |
04:27
π
|
godane |
i'm uploading STart Magazine |
04:27
π
|
godane |
its about the atari st |
04:30
π
|
joepie91 |
db48x: nettalk |
04:30
π
|
joepie91 |
http://www.ntalk.de/Nettalk/en/index.php?page=Screen |
04:30
π
|
joepie91 |
it's awesome, especially for someone used to an IDE |
04:31
π
|
joepie91 |
(and for someone with a lot of channels because multi-line tabs) |
04:31
π
|
joepie91 |
I have it set to English though |
04:32
π
|
joepie91 |
I can't German very well :P |
04:33
π
|
db48x |
ah, I've never used that one |
04:35
π
|
joepie91 |
it's quite awesome |
04:36
π
|
db48x |
I just use erc |
04:37
π
|
joepie91 |
db48x: it has 1. proper server management (much better and simpler than for example xchat or kvirc), 2. multi-line tabs with icons so you can keep an eye on a lot of channels at the same time 3. speed-scroll where you can keep the middle mouse button pressed and move up and down to scroll at really high speed without flicker (very useful for quick scanning of backlogs) 4. mouse gestures 5. autocompletion popups |
04:37
π
|
joepie91 |
(so you can see what you're doing) that work for more than just nicks, basically all commands have full auto-completion 6. command syntax hints 7. optional spellcheck |
04:37
π
|
joepie91 |
and 8. a very nice DCC file transfer manager, but DCC doesn't work under WINE |
04:37
π
|
joepie91 |
so that point is sort of moot |
04:38
π
|
db48x |
cool |
04:38
π
|
joepie91 |
plus in-log search (so not showing matching lines in a separate window, but just moving to the match position in your backlog), a practically infinite backlog, etc |
04:38
π
|
joepie91 |
a backlog for a single channel can probably hold about 150k-200k lines |
04:39
π
|
joepie91 |
if you're a heavy IRC user, nettalk is really awesome for that kind of stuff, haha |
04:39
π
|
db48x |
hmm. erc's buffers are infinite, but channel and server management is terrible |
04:39
π
|
joepie91 |
isn't erc curses though? |
04:40
π
|
joepie91 |
because I've never seen a curses-based IRC client that had proper server/channel management |
04:40
π
|
db48x |
no, it's got whatever capabilities your emacs has |
04:41
π
|
db48x |
if you're running emacs in a terminal, then it's obviously going to use whatever text rendering your terminal does |
04:42
π
|
joepie91 |
ahh |
04:42
π
|
joepie91 |
also, one very important feature that nettalk has that I forgot (though weechat for example has this as well): preview of PMs/highlights in status bar |
04:42
π
|
db48x |
cool |
04:43
π
|
joepie91 |
also this is what network management looks like: http://owely.com/6a3nKN http://owely.com/91iw2AZ http://owely.com/515krZ0 |
04:43
π
|
joepie91 |
(that's not a full screenshot, I decided to spare you the sight of 5 rows of tabs :) |
04:44
π
|
db48x |
heh |
04:44
π
|
joepie91 |
(the $join commands are run-once on-connect commands... they are automatically placed there when you disconnect so that it auto-rejoins on the next connect, and it will forget about them afterwards) |
04:44
π
|
joepie91 |
basically, it auto-remembers your channels, servers, etc |
04:45
π
|
joepie91 |
it's quite impressive and I'm puzzled as to why xchat doesn't have this yet |
04:45
π
|
db48x |
nice. erc definitely could use some work in that area. |
04:45
π
|
joepie91 |
most clients could, actually |
04:45
π
|
joepie91 |
I can't really recall a client off the top of my head other than nettalk that, by default, remembers your servers and channels |
04:45
π
|
joepie91 |
sure, probably erc/irssi/weechat/kvirc can be *made* to do it |
04:45
π
|
joepie91 |
but not by default |
04:46
π
|
joepie91 |
xchat maybe as well, but that would probably be plugin hook hell |
09:10
π
|
Smiley |
sigh |
09:10
π
|
Smiley |
what the fuck |
09:10
π
|
Smiley |
http://www.twitlonger.com/show/kngs7f |
09:10
π
|
Smiley |
grow a pair. |
09:10
π
|
Smiley |
: |
09:12
π
|
Smiley |
auto saving of channels/networks? |
09:12
π
|
Smiley |
that would be nice |
09:12
π
|
Smiley |
irssi remembers what windows were doing things... |
09:12
π
|
Smiley |
but doesn't reconnectr :S |
09:34
π
|
joepie91 |
Smiley: there's an irssi plugin for it |
09:34
π
|
joepie91 |
iirc |
09:34
π
|
joepie91 |
at least, that's what someone claimed to me at some point |
09:41
π
|
Smiley |
:o |
09:42
π
|
Smiley |
tbh I've never looked, and I rarely join new channels |
10:06
π
|
SketchCow |
balrog_: What's the channel for discferret again? |
10:14
π
|
Smiley |
https://gist.github.com/4520930 may come in useful in the future if some site decides to try and stop us via capcha. |
10:15
π
|
Smiley |
though he's stated the obvious solution without thinking about it |
10:15
π
|
Smiley |
Slide all away to the left |
10:15
π
|
Smiley |
slide right and check via audio. Long audio == keep going; else correct |
10:33
π
|
ersi |
Smiley: Configure your irssi ;p |
10:33
π
|
ersi |
networks, channels |
10:34
π
|
Smiley |
it is all configured. |
10:34
π
|
Smiley |
just new channelgs... |
10:37
π
|
Smiley |
hmm |
10:37
π
|
Smiley |
i wonder how high up my house is |
13:16
π
|
ersi |
http://i.imgur.com/7MYXE.jpg |
13:17
π
|
db48x |
nice |
13:18
π
|
ersi |
http://imgur.com/gallery/i0jVO |
13:18
π
|
ersi |
hurrrrrrrr |
13:19
π
|
ersi |
Ooh ooh, highly related to ArchiveTeam: http://i.imgur.com/kaHwQ.jpg |
13:30
π
|
db48x |
lol |
13:46
π
|
db48x |
hmm, something may be wrong with this computer |
13:46
π
|
db48x |
I can see it draw every row of pixels for every letter it puts up on the screen individually |
13:46
π
|
db48x |
it's booting, so that's the bios drawing things |
13:47
π
|
balrog_ |
SketchCow: #discferret on freenode. I'll poke philpem later |
13:47
π
|
balrog_ |
it's been rather slow there for a few reasons, mainly because philpem has been dealing with personal issues (the worst of which seem to be resolved) and I have been doing chip decamping and other stuff like that |
13:47
π
|
Smiley |
db48x: slow monitor |
13:48
π
|
balrog_ |
we got a lot done in August to September of last year though |
13:51
π
|
db48x |
Smiley: heh, it's a crt |
13:54
π
|
Smiley |
the,..... magnety bit is going? |
13:54
π
|
Smiley |
is it whining? |
14:02
π
|
db48x |
no, the monitor works fine with my other computers |
14:09
π
|
balrog_ |
are you using the proper refresh rate? |
14:09
π
|
Smiley |
possibly confused bios |
14:11
π
|
db48x |
it cannot possibly be the monitor |
14:11
π
|
db48x |
it's literally taking so long to update the frame buffer that it takes seconds to draw each character |
14:12
π
|
db48x |
if the monitor were only drawing 8 frames per second, it would be invisibly dark |
14:12
π
|
Smiley |
lol |
14:12
π
|
db48x |
it's a crt, the persistance time is a few ms |
14:13
π
|
db48x |
and if the refresh rate were really that low, the updates between refreshes would be bigger |
14:13
π
|
db48x |
it's not like there's a buffer between them |
14:14
π
|
db48x |
probably the bios just didn't initialize the cpu correctly, so it was running at some insanely low rate |
14:15
π
|
db48x |
reset button fixed it |
14:15
π
|
db48x |
worrysome though |
14:15
π
|
Smiley |
also funny ;) |
14:19
π
|
balrog_ |
wrong refresh rate might screw up the circuitry in the CRT |