October 22, 2017, 03:13:39 am
News:
Pages: [1]
Print
Author Topic: Grease Monkey/userscripts for Neuros.TV?  (Read 9434 times)
jsdf
Sr. Member
****
Posts: 329


View Profile
« on: March 05, 2009, 03:51:35 pm »

Hi everyone,
We've been thinking a lot about how to get our users of Neuros.TV more involved in creating 'plugins' for the site.
For example, wouldn't it be great if you could add a Neuros.TV application that lets you "Find songs for this show" on any show in the system?  There could be a collection of these scripts that users with limited knowledge of Javascript/DOM could build.  Neuros can host these and users can add them seamlessly from their LINK.

If this sounds like Greasemonkey, that's because that is where the idea came from.  The difference between Greasemonkey and what I am thinking is that we want to constrain the area in which scripts are allowed to modify (at least initially, until we find a good way to scale it up).  We want additions to the functionality of Neuros.TV, not wholesale changes.

Thoughts on this?
Anyone have experience with Greasemonkey scripting and its underlying code?
We'd love your feedback.
Logged
bagster
Sr. Member
****
Posts: 270


View Profile
« Reply #1 on: March 05, 2009, 09:51:51 pm »

I was playing pmog few months ago, and I remember 2/3 players that were good with greasemonkey.
I'll try to remember their name and contact them.
Logged
flinkisme
Jr. Member
**
Posts: 60


View Profile
« Reply #2 on: March 06, 2009, 11:35:17 am »

I'm not so sure it'd be cost effective to replicate the functionality of greasemonkey.

If Neuros provided a dom doc with suggested areas for tinkering, that would go a long way towards channeling development in the direction you'd like. I doubt it's possible to stop greasemonkey scripts from making changes as the programmer heart desires.
Logged
jsdf
Sr. Member
****
Posts: 329


View Profile
« Reply #3 on: March 06, 2009, 03:02:15 pm »

flinkisme,
Then, using greasemonkey as our widget engine is definitely worrisome.  Sure, some cool stuff can come out of it, but I hesitate on wholesale changes to Neuros.TV coming out of specific scripts, at least through sanctioned means.  (For those who want to install greasemonkey, they absolutely could and we will provide a DOM doc).

For something simpler and more controlled, what about a system whereby:
1.) Users can pick their plugins from Neuros.TV
2.) On page load, the system calls out to specific plugins' registered URLs via AJAX queries, passing to them vital information (show, season, episode, etc.)
3.) Plugin returns code to format itself in the page and provide interactivity.

The hardest part to me is making sure #3 doesn't show malicious code...

Is this a reasonable way to go, or should the Greasemonkey fully open approach be the way to go?  Thoughts?
Logged
flinkisme
Jr. Member
**
Posts: 60


View Profile
« Reply #4 on: March 06, 2009, 03:27:11 pm »

Well, using greasemonkey as the engine alleviates the need for creating one from scratch.

You could provide scripts for Neuros.tv directly from Neurostechnology.com or from Neuros.tv. I think I'd add them somewhere on the support pages.

If you host them, then you have the opportunity to vet the code as it is submitted. I understand it's a worrisome things to try out new code. I always spend some time to make certain what any script is doing. To the best of my knowledge, greasmonkey will soon become part of the firefox core. Right now, if you click a link to install a script into greasemonkey, the engine provides you an opportunity to check the code before doing so.

If the code is vetted by Neuros then some of the worry for non-techie users is lifted. Most of the worry with the scripts is for cookie/password theft. When looking through the code you can ensure that the script contains no foreign (non-neuros) URLs.

Greasemonkey scripts only function on URLs specified in the script's header or on URLs manually added by the user.

A greasemonkey script can easily do everything you described. I don't know about a plugin to a greasemonkey script, though. I think it would simply be an additional script designed to work with other scripts for neuros TV.
Logged
infrared
Full Member
***
Posts: 182


View Profile
« Reply #5 on: March 06, 2009, 04:43:46 pm »

I would love to see an API exposing app.neuros.tv... just a couple database hooks (maybe a query string in the URI) I will check out greasemonkey but I am more interested in the web server idea for the frontend on the LINK. LINK as LAMP has much potential I believe.
Logged
flinkisme
Jr. Member
**
Posts: 60


View Profile
« Reply #6 on: March 07, 2009, 03:56:11 pm »

LAMP is great, but I'd rather stay low on RAM usage. Adding Apache and PHP is going to suck up a fair bit of RAM. Not to mention that you'd still need a web client to access it.
Logged
infrared
Full Member
***
Posts: 182


View Profile
« Reply #7 on: March 08, 2009, 12:45:07 pm »

FF is already installed so no web client is needed. Apache does not have to be the web server, although it is open source and highly configurable. There is lite versions of apache (Ive ran it on a nokia phone to publish the contents of the photo folder), so memory is not a problem as long as you limit clients to a handful. PHP is very lite as well and apache would not be needed to process too many server side scripts (on the LINK). It would not be hard to implement a few simple notification services in mysql to sync content, or do whatever, thats user defined from app.neuros.tv. It would also be cool to placeshift my home network content through the LINK as a gateway and uPnP server. That would take some work to get working though. The db would make pushing content to the LINK very easy and make for a more fluid user experience.

I believe this way will lock general users into a browser UI that configures the system (have you ever logged into a cisco/linksys router?) but also allows the user to access the content/services they want. 
Logged
jsdf
Sr. Member
****
Posts: 329


View Profile
« Reply #8 on: March 09, 2009, 11:49:25 am »

Re: LAMP on LINK, this is already in the works, though it will be a very light version.  Specifically, version 1 will be pylon with python, with a simple app daemon in C++/Qt4 using DBUS.

Re: API for app.neuros.tv, infrared, what kinds of things would you want to do with it?

Re: Greasemonkey, flinkisme, good to know it might be standard on Firefox in the future.  And, yes, we could do manual script checking on submission.  I would hate for Neuros to be a real bottleneck, though, as we aren't really staffed for this.  (And even those that are staffed aren't doing so well - look at Apple's iPhone market - 3-6 weeks to get an app published after submission.)
Logged
infrared
Full Member
***
Posts: 182


View Profile
« Reply #9 on: March 09, 2009, 01:12:43 pm »

jsdf,
I want the API so that I can build my own front end. Dont get me wrong I make great use out of  the current app.neuros.tv but I have no control on what is returned to my browser and how it is displayed. I really feel that an API and a user base that supports it will keep this project in line with the spirit of open source. If we have no control over the service that supposed to serve us, we (only speaking for myself but you get the point) will look for (build ourselves) an alternative.
Logged
flinkisme
Jr. Member
**
Posts: 60


View Profile
« Reply #10 on: March 09, 2009, 04:29:41 pm »


Re: Greasemonkey, flinkisme, good to know it might be standard on Firefox in the future.  And, yes, we could do manual script checking on submission.  I would hate for Neuros to be a real bottleneck, though, as we aren't really staffed for this.  (And even those that are staffed aren't doing so well - look at Apple's iPhone market - 3-6 weeks to get an app published after submission.)


I would like to see you deluged with a few hundred thousand apps, but I'm not going to worry real hard about it happening. I'd be surprised if it took more than a minute per script.
Logged
jsdf
Sr. Member
****
Posts: 329


View Profile
« Reply #11 on: March 09, 2009, 05:23:03 pm »

jsdf,
I want the API so that I can build my own front end. Dont get me wrong I make great use out of  the current app.neuros.tv but I have no control on what is returned to my browser and how it is displayed. I really feel that an API and a user base that supports it will keep this project in line with the spirit of open source. If we have no control over the service that supposed to serve us, we (only speaking for myself but you get the point) will look for (build ourselves) an alternative.

Understood.  We are still pretty actively changing the way the service works, as it is somewhat of a hodgepodge of other services mashed together (to put it non-technically...).  So, while the longer-term goal is to provide an API to our search, in the short-term we still need to find the right way to deliver correct, applicable results to users quickly.  Until we do that, I don't think we have a lot to offer in way of an API.  Mind you, I don't think we are really far off, we just aren't there today.
Logged
infrared
Full Member
***
Posts: 182


View Profile
« Reply #12 on: March 09, 2009, 09:35:00 pm »

jsdf,
I totally understand the logistics behind the scenes you guys are dealing with. I beleive it is like this.... There is two key parts to neuros.tv (1) the database that supplies the links to the content and (2) the presentation layer where the user selects the content to be viewed. correct? What are the future plans to use the user's creditials as I do not see why there is currently a login process except to collect load usages etc.. A user space would be a cool addition i.e. preferences for the searches, thumbnail size, highlight previous viewed toggle, history settings, queue ... etc.

Anyway of getting a mirrored instance of dev.neuros.tv up with ftp access to the http server and a couple stored procedures ? Wink
Logged
jsdf
Sr. Member
****
Posts: 329


View Profile
« Reply #13 on: March 10, 2009, 04:40:52 pm »

I totally understand the logistics behind the scenes you guys are dealing with. I beleive it is like this.... There is two key parts to neuros.tv (1) the database that supplies the links to the content and (2) the presentation layer where the user selects the content to be viewed. correct?

At a high level, yes.  At a more detailed level, our video sources are both integrated third parties (e.g. Hulu, whose show results are very clean on Neuros.TV) and a generic search provider with whom we are working with.  The latter provides a backstop to enable users to do general keyword searches and find videos from providers we haven't yet integrated.  But, because of this, the handling between the two result sets is not as clean as it could be, and a lot happens behind the scenes to make the data a little cleaner.

What are the future plans to use the user's creditials as I do not see why there is currently a login process except to collect load usages etc.. A user space would be a cool addition i.e. preferences for the searches, thumbnail size, highlight previous viewed toggle, history settings, queue ... etc.

That's exactly right - a user space will let users build a "My Shows" page, customize their preferred search providers, and allow them to customize the service in a large variety of different ways.

Anyway of getting a mirrored instance of dev.neuros.tv up with ftp access to the http server and a couple stored procedures ? Wink
We aren't there yet, but hopefully will be soon. Smiley
Logged
VDH45
Newbie
*
Posts: 17


View Profile
« Reply #14 on: July 23, 2009, 05:51:13 pm »

What about XBMC plugin for neuros service? It should be easy to code browsing/searching neuros.tv within XBMC/Boxee video plugin section, if API is available. All the web page navigation user will have to do would be on "final" page of the content provider. And if Neuros can figure out how to switch full-screen automatically, then would be almost seamless user experience.
Logged
Pages: [1]
Print
Jump to: