Another geekspeak post - I really should start using tags/labels....
So something occurred to me today... Given that Google Chrome uses it's preferences file for just about everything, and the thing that I deleted which people seemed to be panicking over are essentially bits pieces in preferences (except in ChromeOS), could this all be managed?
On the ground, this seems to me that it would be easy. JSON, the format Google Chrome uses to store settings, is a machine thing. Sure, you can edit it. And you can even see sense in. But it is essentially a machine readable format (it's not all that flexible in terms of syntax. Look at where you do and don't place a ',' for example). Using python, it's fairly trivial.
But dealing to the master_preferences file only really deals with the initial deployment. What happens when you want to do something like block an extension after a user has run Chrome for the first time?
So I'm thinking, you could use a wrapper. A script that eventually launches Google Chrome, but can do a few things before hand. The easy answer would be to make it remove extensions. But it could go beyond that. It could alter the preferences file... which means that you could have a web based tool that can set preferences, which then gets injected into the user's preference file just before they launch Chrome....
You'd have to have a wrapper on the client machine. Once the client's there though... how would you limit the control a server has? Should it look at domain name? Or do you rely on the client knowing where to look for settings (something it'd have to do anyway but could be construed as a giant security issue)?
Something else also occurred to me. If I were to do a tool for generating a master_preferences file, and it was usable over the Internet, could I make it bring in some money via advertising? Although it's something I'd loathe, and there's a very good chance that I would offer up the source code (without advertising), is it a way of generating a bit of an income?