FileNotFoundStorage.com Design Document
-
I'll take a look when I get home tonight, sounds like fun
-
-
Nice, the double whammy of language and host provider.
I'm going with c# .net and godaddy though. Make of that what you will.
-
On second thought, there's no reason I can't use both languages.
-
ASP.NET on GoDaddy Windows hosting? There's TRWTF.
-
SECURE file hosting, Keith!
-
Actual retrieval is out of scope (per first post) but we can use some of these ideas for "finding" files.
I think I should make refresh take you to random areas on the page or modify your search results.
Maybe modify a files name encryption if it shows up in search results?
-
API... There's another area ripe for abuse...
API - standing for All Purpose Intercourse.
-
Nice, the double whammy of language and host provider.
I'm going with c# .net and godaddy though. Make of that what you will.
choosing c# over django is insulting to Guido.
-
I really love the file sharing via short url.
-
-
SECURE PERFORMANT storage, ARANTOR!
-
-
Django Unchained!
There really needs to be a language called Tarentino, which handles thread closure... brutally. It should have a language construct that uses 'does' instead of 'if' so that we can check whether
does (marcellusWallace.lookLikeABitch())
-
-
That's ok. I never did like the guado in ff10
-
This whole project is intended to waste time, kind of like this forum.
-
Then you have a perfect contributor in @Nagesh.
-
This notification is tricky. I am getting it again and again. What's up with that?
Ah, the "Braillant @Nagesh Bean"!
-
This notification is tricky. I am getting it again and again. What's up with that?
Probably nothing - I seem to remember someone observing that you get a notification for every time a user first mentions you, but that all notifications that originate in the same topic somehow all refer to the same, wlthough arbitrary, post.
-
The service should provide lengthened URLs, not shortened URLs.
-
I could provide both. But how long is long?
-
I could provide both. But how long is long?
There is only one way to find out...
-
For Science!
Ok, looks like Discourse has a problem with lengthy urls...Edit: confirm, it's Discourse. The url itself triggers a 404 as expected.
Filed under: yes, could look better, but I don't have the time right now
-
It's forcing the clients to generate PDF to send requests to the server that elevates the wooden table count…
Send requests to the server? That would bog down your server. The server should send the hash to the client and the client side can run the unhashing. That way you're ALWAYS first in queue because you are your own queue!
If enough people signed up, you could build the world's greatest hash rainbow.
-
Fucking hell, @blakeyrat was right, 99% of all topics devolve to Discurse bug hunt and discussion.
I mean... ummm... no, I never saw him, what are you talking about?
-
Hard to say without seeing how long a URL that's been neither shortened or lengthened would be - just needs to be longer.
Generate 3 different hashes of the file name and concatenate them, or something.
-
Generate 3 different hashes of the file name and concatenate them, or something.
Too easy to guess! I suggest Hamming code it, but instead of parity ASCII codes have to add up to 83.
Filed under: For security!
-
1 short of two meanings of life, I like it.
-
1 short of two meanings of life, I like it.
It's not short of it. It's just zero-indexed.
Filed under: Totally the intention.... yup!, Spell checker doesn't work in this browser, bet you can tell
-
Generate 3 different hashes of the file name and concatenate them, or something.
I vote for concatenating the hashes of each individual character, salting each hash with the file content hash.
Hash.
-
Some thread somewhere that I can't bother going to find due to discourse suck search
Here:
I could provide both. But how long is long?
Old but apparently still applies:
-
This could work. Keep in mind, accessing this URL won't actually go to a valid file, it will always ultimately go to some type of 'Oops!' screen since the file will get renamed.
-
-
I cheat. I don't use Discourse search.
-
@PJH
New badge recommendationCheats on Discourse
-
I hate to be a party pooper, but the idea of "write-only file storage" has been done at least twice:
People will find the website, read it for a minute before figuring what's going on, chuckle, then move on to something else.
If you want a joke that will really make an impression, name it "Super PAAS Enterprise Storage Inc" (or something that sounds just believable enough) and actually works as a file storage, but is riddled with WTFs like (as mentioned) requiring you to do all the communication as encoded PDF files, or ROT-13 "encrypting" all text in the stored files for extra security, or having a "sync" client that only supports absolute paths and syncs C:\Windows\ by default....
Then some disgruntled employee somewhere can convince their technologically illiterate boss that it's better than Amazon S3 and that he'll handle the entire migration, just before resigning from IT...
P.S. I suggest using ASUS WebStorage as a backend for extremely slow transfers.
-
Like the first couple of bits of the file.
-
That's fine that it's been done before, I'm sure neither one will be as securetm or as performanttm or as transienttm as this service will be.
Plus, it's just for shits and giggles, neither one has as awesome of a name as filenotfoundstorage.com
-
Old but apparently still applies:
I think I tried > 2400 characters in the link in this post; Discourse can't resolve the link but the url alone works in Firefox.
-
Anybody want to kick in for 404.solutions? lol
-
The server and client should clearly be communicating by sending programs generated on the fly in a custom scripting language that only they understand. The scripting language will, for style, use a 2d setup not dissimilar to piet.
For security purposes, a polymorph engine will run on the generated program to complicate
debuggingreverse engineering.And of course, the scripting language needs to be missing critical features, so that they have to be implemented manually each time. And it is designed to make self modifying code required. Want to loop? Edit the access command! Oh, and the loop end check too.
Finally, the client is mostly implemented using the same language and loads received commands in the same address space as the client code. And the received commands "accidentally" trample over parts of the client, causing the entire thing to spiral out of control if even the slightest error happens.
-
And your proposal for implementing this feature in a weekend?
-
Pizza and your favorite beverage!
-
custom scripting language that only they understand
Encoded in UTF-32; keywords and identifiers use only code points between U+1F000 and U+1F02F, inclusive.
-
@faoileag inadvertently reminded me by posting in a topic I CBA to find now because Discourse breaks "Duplicate tab" functionality by fucking with the history, that we need a mobile app, stat!
I'm torn between making a native app just embedding some HTML for each platform separately, or finding the most WTF-y framework to do it for us.
Proper native apps are, of course, out of the question.
-
I was about to scream "Nooooooo!", then I read the bit where you said that you wanted to do it. It's ok then.
-
use a 2d setup not dissimilar to piet
Like Funge-98, incidentally my favourite programming language family. Self-modifying, available in 1, 2 or 3 dimensions and fully extensible to implement external functions like raw network access.
-
Yeah, that's never going to happen.
-
You clearly have never experienced the FUN of implementing new algorithms in 2 dimensions!