@J.Starnes did you actually type all that text from the image to be text? --> askubuntu.com/q/982304/522934

@Videonauth No like the edit summary noted it was ocr then manual corrections and formating.

@J.Starnes ah ok which ocr are you using for that?

text fairy on android. gocr on laptop or desktop. Both have annoying error rates but manual correction is not to bad.

i see, sad this hasnt an GUI and no abilities to learn from failures to increase to detection rate

I usually edit the file in gedit to use the regex find/replace function to fix similar errors

figured that :) still i will later have a look for good ocr solutions and let you know if i find any

but this is already better than typing all that

great ocr and music ocr would be great, but dosen't seem to be a high priority. Or perhaps its a significant lack or suitably skilled developers.

[ SmokeDetector | MS ] URL-only title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, +3 more: nutritionplanreview.com/male-onyx/ by user766823 on askubuntu.com

@J.Starnes well it is hard to create an intelligent pattern recognition which is able to learn from mistakes, even on windows there is only one company actually which got this right, but then it still needs long time to be able to parse hand written stuff etc.

[ SmokeDetector | MS ] URL-only title, bad NS for domain in body, bad NS for domain in title, bad keyword in body, bad keyword in title, +6 more: getnutritionshelp.com/trembolex-ultra/ by Compromed on askubuntu.com

@Videonauth ocrfeeder can serve as a gui for gocr or other ocr back end. Mostly it helps with layout. It does allow for side by side comparison of the scan and text.

I found three other ocr softwares in the archives tho, tesseract, cuneiform and lios, but no clue if they are better

Good morning Zanna

@Zanna I'm almost convinced that this is the right answer, but I'm not sure how the second answer there comes into play, might be important too, and to be honest i have had now over an hour to ponder over it and came to no conclusion (regarding this)

@Videonauth @J.Starnes would you prefer your conversation be moved to AUGR or to the Island?

Island

@Videonauth I tend to think the other answer is an extra step, but it might be an alternative method!

@Zanna I tend to assume the first of your thoughts, but i could not be sure.

@Zanna that conversation doesn't seem to be off-topic actually - sorry folks!

:P

Yes it was about how to make a quick transfer from image text to normal text in case of editing

if we would find a good OCR solution this would save us many comments ala 'Do not post images of text'

but I think we should keep asking people to do it the right way too...

true

its just, I for myself would not sit down to transfer picture to text, but would be more inclined if i could use an ocr solution for this

that's frustrating

why?

I don't think it's a dupe of that and I don't want it to get deleted because the answer is good (and would make no sense whatsoever if it were merged to the supposed dupe!)

@Zanna well i just gave it me reopen vote, guessing the other one is from you

thanks!

if the others step in too we should be able to open this without the review queue

sighs deeply: Another one of those, but you just told me that this is a bad idea i want to do it never the less ...

This question was double posted and both questions are duplicates of: askubuntu.com/questions/961967/….

@karel plugged in the second dupe target..

This should still be closed as a duplicate.

@Videonauth Did you test that? If that's how it works, I am surprised. My understanding was that the user's password decrypts the key which in turn decrypts the encrypted home directory. I think the answer is actually reasonable, and I would at least remove my downvote and possibly even upvote it, if it better stated when it does and doesn't give meaningful security.

@EliahKagan not had the calm yet to setup a VM to test

and did you see from yesterday chat.stackexchange.com/transcript/message/41477305#41477305

Even as it stands now, I think that answer about encrypted home directories is far better than the one that says to give users just the ability to execute specific commands, doesn't say what commands would be safe, doesn't explain how to figure it out, and links to two answers on other SE sites that show commands that aren't safe.

You'd pinged me earlier here about that; sorry I hadn't replied. Although your commented suggestions were helpful, I don't think they address the core flaw in the post. I actually downvoted the two answers it linked to on other sites, which was an extremely easy decision for me: those answers recommended insecure, dangerous crap, with meaningful caveats only in comments written by others.

It was a slightly harder decision for me to downvote that answer, though I did do so. Even if it quoted the wrong examples from the linked answers on other sites, they're still wrong examples, and the answer as a whole would still be dangerous and effectively wrong.

I'm not saying you should do anything further. I really just don't think that answer can be fixed. There's no sign in the question that it would be necessary or useful for the OP to give the other user even limited sudo rights, and in the absence of specific information about what powers the other user needs to have, it would be hard for the answer to become specific enough in a way that would be helpful.

In contrast, the answer about encrypted home directories probably can be fixed. At least if I am correct about how the keys are stored.

@EliahKagan That whole Q&a is a firehole tbh, i deliberatly answered so vague withou going into deteils, in my opinion only a full complete package of the following will work, proper user management, home encryption, and drive encryption coupled with password for root and hardened security etc

There are so many aspects to cover if you want to go really into detail how to harden your linux system that it would bust the limit you have on characters

in fact you could write entire books out of this

Oh, well I have no objection whatsoever to your answer. The question is answerable and not appropriate for closure, but it's pretty general, because the OP didn't state what powers the other users needed to have. Given that, your answer makes a good bit of sense. If you do end up checking how the keys are stored when home directory encryption is used, though, then I think that would be valuable for evaluating and likely for improving that answer.

will do this weekend, have this on my list

just fighting with a new medication a bit

already the past days

also on my list, is maybe posting my python code i made for that answer on peer review to get some opinions on it

Back in 2014, I had wrongly thought this post really was an answer, and it's likely that the reason it wasn't deleted was that I disputed the flags. (Back then, before the new review system was in place, the 10k tools used a system where you could add your flag, or counterflag, posts that had been flagged NAA.) Only the first comment was present, and the problem it describes is inadequate to delete the post.

Still, I should absolutely have seen the problem back then. The software they are recommending isn't for Ubuntu and there's no reason to think it can be run in Ubuntu. So I think I was wrong back then. I've flagged the post NAA. I'm hoping my error can be corrected, even if it's a few years late. :)

@Videonauth That's not a bad idea. I plan to get around to reading through it soon, as you had requested, but there are people there who know Python way better than I do. ...I've never posted on Code Review. Sometimes I see unchallenged falsehoods in high-scoring answers; and I recommend reading them with a critical eye, but that's no reason not to post. A lot of the advice that is given on that site is just 100% excellent and a lot of important problems in code get correctly identified there.

(Perhaps similarly, it is possible for me to make mistakes and say wrong things about people's code, too! :)

well i would count myself as a late beginner - mid novice in python as i do this only hobby wise, i usually get my code to do what i want, if that counts as something

One of the nice things about Python--and I really do think it's a nice thing--is that a lot of the stylistic controversies in other programming languages that tend to generate misinformation are simply dictated, or nearly so, by the syntax of the language. So, for example, you don't see people debating about what brace styles are "objectively correct" in Python, and aside from the everlasting tabs vs. spaces controversy, you don't see people debating indentation style either.

yep since only spaces count in python afaik

My guess is that the result, for peer review--especially the kind of peer review that is given by strangers over the Internet where there is no established corporate or projectwide coding standard so people just use their preferences--is that Python advice will tend to be based more on objective facts than, say, advice about how to code in C.

guess so too

@Videonauth Python accepts indentation by tabs.

oh ok, well i alsways use 4 spaces as tab conversion anyways

ek@Io:~$ python3 -c $'if True:\n\tprint("First line.")\n\tprint("Second line.")\nprint("Third line.")'
First line.
Second line.
Third line.
ek@Io:~$ python3 -c $'if False:\n\tprint("First line.")\n\tprint("Second line.")\nprint("Third line.")'
Third line.

@Videonauth Yeah, so do I.

IMHO its more secure when it comes passing code to other people, if they have different tab size settings the code wich looks wonderfully formatted on my editor might look like a mess on theirs, using spaces is more reliable in that case

so people see code as i wrote it and inteded it to look

Well, I think SE converts tabs to four spaces when they are formatted as code, anyway. So when they copy, they'll get spaces. This is usually fine but actually annoying for makefiles, since (by default) makefiles cannot be indented with spaces.

CV needed: askubuntu.com/questions/632773/…

TIL