Web-based Random Generators: The Future
My god, there are a lot of random generators out there now. Band names, insults, British town names (!?). There are portals with nothing but random generators now.
My poor old Random Logline Generator is feeling a bit under-appreciated. Been wanting to bring it into the modern world for some time now. I've been thinking of the ideal system for generators (including the RLG) since 1996 or so. Though I do admit sometimes too much grand thinking gets in the way of actually implementing things (Project Xanadu, anyone?).
Still, the random generators out there are fairly primitive (even compared to mine in some ways, which is written in old-school Perl). Sure, they're much prettier, written in PHP or Javascript, and some have extras like letting users submit words and their favorite results. But behind the scenes, things have not progressed much. For example:
Inflexible (Fixed) Grammar Sequences
Most of these appear to be using fixed patterns, typically just one to make a result. RLG's backend can support dynamic sequences.
Proprietary & Closed
While some of these newer generators can be made into widgets usable on various platforms (blogs, Facebook, etc), there's no standard way to grab results or portions of the results to make your own mash-ups of random things across sites.
No Editing of Results
All of the sites I've seen, what you get back is it. You want something else that's similar to what you got, but maybe you don't want "sexy accountants", you want something else but keep the rest? Not possible. Your only choice is to shuffle everything, getting a totally different result. (RLG had a klunky interface to do editing but I took down because few could understand how to use it.)
Here's what I propose: an open-source approach, akin to RSS and Atom, for getting randomized content at the atomic (word or image) level and at a sequence level, from multiple sources.
In plain English, this would enable the ultimate hat full of cards. Each "card" deck classified by what kind of thing you wanted, and as big as all the cards your "friends" or you had in the deck when you asked for it.
So if someone wanted to build a random anything generator client, they would make calls to the API requesting the tagged thing itself (e.g "occupation") and would be able to include their own lists of things, optionally making those lists available to everyone else too via subscription.
The interface should be RESTful, offer both JSON and XML results, and anyone with a web server should be able to host lists.
However it works, it needs to be fast enough to aggregate the results from all the sources, while maintaining the information needed to shuffle any subset of the results.
How about it, Google? Yahoo!? Anyone?
My poor old Random Logline Generator is feeling a bit under-appreciated. Been wanting to bring it into the modern world for some time now. I've been thinking of the ideal system for generators (including the RLG) since 1996 or so. Though I do admit sometimes too much grand thinking gets in the way of actually implementing things (Project Xanadu, anyone?).
Still, the random generators out there are fairly primitive (even compared to mine in some ways, which is written in old-school Perl). Sure, they're much prettier, written in PHP or Javascript, and some have extras like letting users submit words and their favorite results. But behind the scenes, things have not progressed much. For example:
Inflexible (Fixed) Grammar Sequences
Most of these appear to be using fixed patterns, typically just one to make a result. RLG's backend can support dynamic sequences.
Proprietary & Closed
While some of these newer generators can be made into widgets usable on various platforms (blogs, Facebook, etc), there's no standard way to grab results or portions of the results to make your own mash-ups of random things across sites.
No Editing of Results
All of the sites I've seen, what you get back is it. You want something else that's similar to what you got, but maybe you don't want "sexy accountants", you want something else but keep the rest? Not possible. Your only choice is to shuffle everything, getting a totally different result. (RLG had a klunky interface to do editing but I took down because few could understand how to use it.)
Here's what I propose: an open-source approach, akin to RSS and Atom, for getting randomized content at the atomic (word or image) level and at a sequence level, from multiple sources.
In plain English, this would enable the ultimate hat full of cards. Each "card" deck classified by what kind of thing you wanted, and as big as all the cards your "friends" or you had in the deck when you asked for it.
So if someone wanted to build a random anything generator client, they would make calls to the API requesting the tagged thing itself (e.g "occupation") and would be able to include their own lists of things, optionally making those lists available to everyone else too via subscription.
The interface should be RESTful, offer both JSON and XML results, and anyone with a web server should be able to host lists.
However it works, it needs to be fast enough to aggregate the results from all the sources, while maintaining the information needed to shuffle any subset of the results.
How about it, Google? Yahoo!? Anyone?
Labels: creativity, google, programming, random generator
1 Comments:
I love your Random Log Line generator and have blogged about it more then once. It is a great way to come up with challenging starting seeds.
Post a Comment
<< Home