Hacker News Reader

Comment by dalke | Hacker News Reader

Ideas are plentiful. Your's comes across similar to "the next generation of the social web"; that is, "Automator on some drug that made it crazy fast" has nothing in the way of how such a thing might work. Even after thinking about the problem a lot over the years, I don't see how one might build a "crazy-fast Automator" along the lines of what you propose.

Given your background, you likely don't either. You come close to the discussion from http://news.ycombinator.com/item?id=2579668 of "Idea Guy Looking for Developer." Read the linked essay and the responses.

BTW, there are other drugs in Greenland. The average inhalant use is higher there than the rest of Europe, says http://www.espad.org/greenland .

View on HN

What is inhalants even? People here don't really hide if they do drugs, believe me. Cannabis is illegal but even police knows who does what, and don't actually do much about it because pushers will beat up anyone trying to sell anything but cannabis, and crime rates rise should any pusher be arrested because the kingpins will withdraw the drug from the market leaving a big demand with very little supply, price rise and people will have to steal things in order to afford the drugs.

Crazy-fast was bad choice of words on my behalf, I ment something more like crazy-efficient but I guess no drug makes you that.

I can give an part example of things that could improve workflow a great deal for almost anyone with a client that approves graphic work in lets say Photoshop. The tasks that is only related to the stuff you do on the computer would be: making a new document, naming it, (if you work with both print and web you might want different units, rulers and color profiles for either type), making the graphic work, saving it some place associated with the client, exporting it for web in some reduced resolution that is big enough for the client to see the details good enough, opening up your email client or webmail, adding in the clients contact information, maybe you have some form of request tracker that needs an ID field in the email, attaching he file for approval, writing a message for your client, press send. Turning all those steps into something more enjoyable like: selecting the client in a dashboard, give the task a name, make the graphic work, save, write a happy message for your client, press send.

Adobe allows JS to be used for improving workflow so it would be ideal coming up with a framework that would use JS. Web developers and even most designers already know JS which makes it an ideal choice.

How often do you name a file according to a name convention? Whatever text field you enter that name in should just allow you to type the filename not having to think about conventions. If the convention only allows a typical lowercase a-z, 0-9 and dashes only, then the field you type in the name in should convert all characters as you type to lowercase, spaces could even automatically turn into dashes, international characters like å, æ, î, é should perhaps turn into their a-z equivalents, and everything else should be ignored. No errors, no interruptions, just work like you like it.

Making an iOS app for both iPhone 4, older iPhones and iPad, you'd have to save the (most likely) same icons and so forth in different resolutions with same prefix but different file naming that allows Xcode to recognize the different versions for the different targets.

For every website project you have a starter pack that consists of a bunch of different things, like a css-reset, maybe jQuery, maybe Modernizr, maybe a responsive mobile-first css framework... Downloading to update all those could be a task in it self...

You like a certain folder structure maybe but the developer you're gonna pass the work on to for coding the project is really strict about how they should be delivered. The developer could give you a profile-file which the workflow framework would use to run a preparation check of the files, checking images are in correct format, packaging your work so it's ready for the developer.

You use Photoshop for something that has to be both printed and for a web campaign, there should be a switch that would make sure stuff like black in the CMYK version is 100% black. You share your Photoshop files between computers? Nothing is more important than having to send fonts back and forth should the other computer not have the font you're using.

You've tried spending time trying to figure out why some server-side code is not working as expected only to realize that there is some cached stuff somewhere that is teasing you. You just saved your changes to the server-side code, you alt+tab, the browser should flush the cache and reload the page when ready.

Your website project is ready for production? Minify all the stuff, in some cases you'll merge files, gzip them, ou get the drill...

In the end, less frustration, more getting done what you like. Computers are supposed to make the tasks easier, not get in our way.

http://en.wikipedia.org/wiki/Inhalant_abuse . "Huffing", "sniffing glue", etc.

The tasks you are talking require massive customization. It's a lot of niche development; few clients in each niche and a lot of work => more like consulting work than product development.

Niche customization is the reason why VB, Automator, etc. were all developed. You also have a massive sales problem - how do people find out about your software?

Consider "website projects". RoR and Django projects each have their own layouts and requirements. Most people using Modenizr are going to be the types of people to mix-and-match, and not want to use some tool to select the parts.

Then you switch to talking about xcode, and Photoshop. And server-site code. And minification. Each of those is hard in their own right. Then you'll find that someone really wants the filename name to be "España.jpg"

The dream of making it all work together smoothly is nice, but there's no easy route to there from here. Nor have you outlined what that route would look like. You've only stated the goal and have some idea of the tools you might use for that goal. Eg, how does one actually specify the "folder structure" and content validators? Is it a GUI? A configuration file? How does one handle cache flushing, including configuration information for the different deployment environments, connections to the different cache layers, and timestamps placed in GET requests to force new fetches?

Solve one of those problems and you'll have people using it. Don't try to solve the entire problem - I suspect there is no general solution.

I realize it'll take a huge effort. I don't believe making the "framework" is consulting work. I don't mind giving the framework away for free with some common/possibly popular workflow schemes, showing it's potential and then whoever wants to can be hired for developing schemes for those who can't make them themselves. Or another more likely scenario, people could make their own schemes fitted for themselves, giving them the option to share it with others if they want to.

The framework would be modular. Sure, no two projects are actually exactly the same but there is a whole lot of overlapping, repetitive, unnecessary, tasks.

I'm not saying you lock into forever using whatever name convention you're using.

It's my first time ever proposing to work with strangers on the internet. I don't have a whole lot of experience with it so I can't really come out with every angle of the final product before anyone asks. (You might think you're asking but you're not, you're saying "You are an incompetent "idea man" who just came up with something you assume is unique and now you want a developer as your pet slave". Instead of saying things like "Nor have you outlined...", you could say "But how do you plan to deal with...").

I imagine folder structure being defined in Finder, the whole concept of the framework is based on the idea that all apps are great at their own thing, so Finder being the core of navigating through folders and finding files it would be logical. While the framework running in the background, Finder could get some extra toggle buttons or drop down menu in the toolbar to let you select what the folder should have of attributes "Like this should be where the XX of the project goes". Could use color labels to make it faster to distinguis between the different folders if one didn't find them hideous. Depending on what'd you were developing cache flushing would obviously be handled different, who ever developed the scheme for that workflow would decide how it should be done. Someone likes to do it with changing filenames of the files you've edited. The workflow application would handle this, make sure where ever those files where included in the rest of your project their reference would be updated, and the browser you were using would be redirected to the new url. This would mean that your source files would have to be copied into a the deployment folder.

People are creative and resourceful and if you let them taste something great you should expect some of them to crave for more. I'm not saying the framework should have schemes (one computer is not limited to having a single scheme either) that fit perfectly for everyone but if you'd provide a few and show how smooth it can be I don't doubt that some people will want to roll their own.

Yep, I'm old and jaded. You see me as negative? _shrugs_ . I never said it was impossible, only that it's hard, tedious, boring, and that it goes through well-trodden areas with lots of competition so revenue generation will be hard. Pick one of the sub-tasks and work on it. You'll be happier and with a larger chance of success. Or think of that as a stepping stone to learn more about what's needed for the larger project.

Personally, I think the directory structure layout validation is the most interesting, especially if you can tell it to build your final structure layout. This could include minimization tasks, or automatic image format conversions. Another direction that could go is towards a static site generation tool. But then, I'm a developer.

Your rhetorical style needs work. It's ludicrous to assume that I or anyone else would expect you to have all the details planned beforehand. Neither did I say that you are incompetent. I pointed to that URL for a reason - it's hard to come up with an idea and get developers involved, and that essay gives some of the reasons. Instead, I am pointing out your inexperience and giving advice for the some things you need to distinguish yourself from an "incompetent 'idea man'."

You should also talk a lot less about yourself. Most people don't care enough to read through several paragraphs defending your inexperience to figure out what your point is, the title had little to do with what you wanted from people, and the self-described term "Problem Solver" is another suspicious flag which falls into the "Idea Person" category.

You're right, my rhetorical style is rubbish. :)

I really appreciate you taking the time to help me on my way. I will work on some proof-of-concept apps that can help illustrate the approach I wish to take. Starting with the folder structure layout builder, image format conversion and layout/format validation. Then I'll see if it has something to it.

I'll post when I have some updates. Gotta go.

Reply on news.ycombinator.com