105 comments

[ 117 ms ] story [ 2539 ms ] thread
This is the type of wild stuff I come to HN for... will I ever use it? No. Is it amazing? Eff yes.
If you already have a lot of Java stuff on your resume, why not add it for a bit of humor?

Of course, you'd have to at least get it working and play around.

I bet the author had a lot of fun, but still, I must ask: "why, oh why!!?? Just why!!?"
mspaint is still one of those things I instantly reach for when I've got a framebuffer on my clipboard and I quickly need to make edits for pasting into email/issue/etc.

Almost everyone else on my team uses some other fancy bullshit that is probably better in every way, but I have found deep comfort in the simplicity & speed of this ancient application.

My preference for mspaint is probably out of habit from working for a megacorp that disallowed installing of any third party applications on work computers. It was the most accessible option on any computer in the org.

It's the most accessible on any corporate computer, but also fast to start and fast to use.

I use Greenshot to capture images, but its image editor cannot move rasterized bitmap around... so I still need to go to MS Paint or equivalent.

Interestingly, the Paint 3D that now ships with Windows, is a strangely unfocused app.

It wants to include the poly mesh as a basic data type, like text, bitmaps.

The ecosystem just is not there.

Paint.net is free and gives you photoshop-ish functionality with a more mspaint type of UI. You would probably like it. https://getpaint.net
Same. I use the crap out of MSPaint and Paint.net when I know I will be doing a bit more.

Simple, fast, does just enough.

There is another basic reason and that is bitmaps and text are still crazy useful for basic communications.

I started out using both of those early on in 8 bit land. Still dominant today.

In that sense, MSPaint makes sense as a basic OS tool.

DIWHY

(Incredible talent shown though)

The video where he does search&replace on the images using OCR shows how much effort went into this. Just wild, haha.
Superiority: It's not Eclipse.

Sold.

My curiosity is piqued, but I also found it hard to navigate this web site. How do I see screenshots? When I clicked "Info" on the Theming section I was taken back to the top of the home page

(Firefox, Android.)

Skimming through the demo video on the homepage is rewarding.
This has a vaguely steam punk sensibility of a rethink of those old punch cards using only technology / techniques from the current era. Instead of the cards you've got image files.

Maybe I missed it, but it would be good if it also created image outputs from the program stdout.

Makes me want to take screenshots and paste them as comments in the code, with circles and arrows and stuff.
MS Paint's convenience and simplicity is underrated. I end up subconsciously using it more often than I think I do. Screenshot pastes, quick edits to blur out stuff before sharing etc.
> Screenshot pastes

This is a failure of Windows's screenshot tool though, in my opinion. I find this incredibly clunky. On Plasma/KDE, pressing "print screen" takes a screenshot of the complete screen, and a dialog opens and offers you to copy the screenshot into the clipboard, to save it in a file or to open / edit it with the tool of your choice. You can also take another screenshot of the entire screen, or a window, or an area that you can draw with the mouse, with a delay or not.

On other environments, the screenshot is automatically saved in a file without interaction. I'm not a huge fan but this is what you end up doing most of the time on Windows anyway, but by manually pasting in Paint and saving.

Pre Windows 10 I'd agree. Try Win+Shift+S on Windows 10.
Good to know, what does it do?

Too bad it's not bound to the Print Screen key but I guess it would break many people's workflow.

It allows you to select a region of the screen, then copies it to your clipboard, so you can easily paste it to other programs. It’s the best workflow there is because it requires no other running tools, and the binding is even simpler than ctrl+cmd+shift+4 on MacOS.
The is an option for this in Ease of Use > Keyboard > Print Screen shortcut
Try Windows-Shift-S if you have the Feature Update for Windows 10 20H2.
Win+PrtScr saves the screenshot as a PNG file in Pictures\Screenshots.

Windows also ships with a tool called Snipping tool[0], but that now says that it will be replaced by Snip & Sketch [1].

If you want to quickly take a cropped screenshot you can use Win+Shift+S. It only saves to the clipboard, but a lot of places, like imgur and Github, supports pasting images so I often find it more convenient than to have a bunch of old files cluttering the system.

[0] https://support.microsoft.com/en-us/windows/use-snipping-too...

[1] https://www.microsoft.com/en-us/p/snip-sketch/9mz95kl8mr0l?a...

Snip and sketch allows you to save a copy of your snip. For me, that's the tool that comes up when I do the Win+Shift+S shortcut. Really handy tool and shortcut.
Alternatively, Paint.NET is the go-to for all of my image needs.
It feels like there is a kernel of something realllly cool here but I can't quite tell what it is.
I think the whole idea is cool! It doesn't have to be useful to be cool, in fact some of the coolest projects are eminently unuseful.

I think this is an awesome resume project. They've built an IDE (which is impressive by itself), and then they've got this whole parsing code from OCR thing on top that's not necessarily useful but man is it cool. The fact that it shows a good sense of humor is just icing on the cake.

Like, search and replace on pictures of code contained in PNG files... Just amazing.

Ah I am in no way trying to be dismissive - this is really cool by itself, and it feels like there may be applications of this that would be highly useful at the same time.
That was essentially the goal with this in the end, I was able to talk to red hat developers about the actual IDE part, do a ton of research on the custom OCR, learn about the Google Assistant API, and a ton of more stuff. I'm glad you enjoy the project, I dumped way more time than I probably should have in this silly thing, but it should pay off when I start looking for jobs in the field lol
But does it work with Wine?
Is this meant to be some sort of amusing tool in the spirit of "because we can" or is there a genuine practical purpose to it?
No general purpose. I know the guy who made this and the hype actually started on a Minecraft forum for developers where there were a lot of Eclipse vs. IntelliJ arguments. It was a hilarious addition to any “What is the best IDE?” threads! He became quite interested in OCR and thought this would be a fun project and thus MS Paint IDE was born!
>Can read, parse, and highlight code from purely image files

>Finding and replacing of text from image files

These two features actually seem pretty useful. There's been more than one time i've found code as an image or something it would have been nice to be able to copy.

To be honest, it's come up enough that i've wanted to be able to copy text directly from an image I've wondered if such a tool existed. Never enough to actually go look for one, I'm sure many exist, but enough that i've thought about it at least.

There’s an app called Memos on iOS that has that capability. It’s great for searching images for text, for example I use it to take photos of my receipts, menus, screenshots, and entirely it’s on device and nearly instant (with 40k images). I’m sure other apps exist for various platforms and in the same niche. Looks like iOS 15 is adding text selection to their photos tool so if there’s a phone number visible it’s selectable and interactive.
Not totally related, but enough that I’ll give a shout out to tabula. It can ocr tabular data from a pdf. I’m not affiliated but I really need to get my company to contribute $$$ to the project. For now I’ll namecheck them.

https://tabula.technology/

i thought this was going to generate code based on drawings you made in ms paint.
If you can draw convincing enough text with a paint brush the OCR should pick it up.

But yeah next level would be drawing a square inside of a rectangle to generate “class Square extends Rectangle”

maybe the next level of leetcode will be to write code in this IDE using only the spray can tool, and if you don't get it done in less than 1 minute, they will pass on you. of course, everyone will happily go along with it.
I can finally put MS Paint back on my resume under “technical” skills!

Amazing dedication, made my day!

Making images executable sounds like a bad idea.
I want to see this working in python - good luck reading whitespace in OCR...
Is this more of a...

"Your scientists were so preoccupied with whether or not they could, they didn’t stop to think if they should."

or a

"We choose to go to the Moon not because it is easy but because it is hard."

situation?

Lol. Looks more like the first option to me, but I also feel that there is something very very special about these kind of efforts and people who don’t mind spending their time and energy on things like this, precisely because of the fact that the world is filled with the ‘useful-or-not’ types. In fact, I feel we need a lot more of these ‘do-I-find-it-personally-interesting-or-not’ types, so that the rest of us can keep riding the ‘whether-or-not’ bandwagon and when that one off ‘interesting-or-not’ turns out to be useful, in ways we didn’t foresee, we can ask them to join us.
> if they should

I hear crypto is a good place for this attitude, these days. I actually got an email from a recruiter describing the product as React for smart contracts.

What a time to be alive!
Demonstration of skills demonstrated with a fun example. I'm assuming there is image->Text translation, process automation + functional demo, and UI design.

Digging into the github page it appears a custom OCR solution was created. I wonder if anyone has forked/used this in cool places.

While this is satire, it does raise the question:...

Is it time to consider what benefits more advanced file formats for source code could offer?

Like in this case for example, the image format could allow doodles, notes, and diagrams to be in the margins of the code...

I loved how TempleOS embedded 3D sprites right into the source code!
This man was really smart...
Not only that but you could print the 3D shapes to the terminal and copy/paste them in to other files.
https://youtu.be/mBgIBF9Y6PE?t=485

"We have graphics in the source-code"

I've not seen his livestreams before, but this lived up to expectations. Within two minutes he's gone from demoing some innovative tech to calling people the n-word and spouting conspiracies about the CIA.
YouTube is asking me to provide my ID or credit card to view this -- apparently it's an EU thing ;( Even though I'm pretty sure I watched this video multiple times and have it saved.
It’s so they can feed your info to the CIA.
They already have everything thru my GMail though. I'm more bothered by the fact that for the last 15 (?) years all I had to do was log in and now suddenly it's not enough -- for videos I've already viewed.
That's incredible. I can't imagine how much time it would take one person to make all of that from bare metal.
"I went to a real school and not a CIA brainwashing place"

It's really something to see him demoing something genuinely very impressive while screaming conspiracies.

That is IDE's job because images can just be embedded in markdown comments... Version controlling non-text comments seems to be an unsolved problem though.
An interesting thing to do would be to represent the syntax tree in something other than text so everyone can apply their preferred format.
I read your comment like "This is so crazy, it just might work!"
Looks like its finally time for me to switch to windows...
Jesus what a colossal waste of time
I couldn't put it better myself, 1k+ hours poured into this for a handful of laughs isn't a trade many would do