what will we do with an api

please use this thread to post examples of something(s) you will build on the api. please do not post feature suggestions for the lindens. their only job is to provide a working, hookable interface.

please also offload discussions and hair pulling to other threads. overrunning the general forum with api discussions can only impress cory with how desperately second life needs this.

the first thing i will do is add spellchecking to notecards to aid me when i type howtos, contest rules, project proposals, voting critiques, personal letters, et cetera.

I'll write an AppleScript to log in all my alt accounts every day so I get their dwell, hahaha. Just kidding.

One thing I would do is write something to help edit prims as I build. It takes a stupidly huge amount of work to use LSL scripts to edit objects in-world. If you don't put a script in an object to start with you're borked by the time you realize you need one, and then you need some pretty complicated listener to take the commands. I'd like to be able to script bulk changes to groups of objects, like making all selected objects 10% taller but not wider, aligning objects, aligning textures, etc. This would also be extremely helpful for working with builds that are too large to be linked, or doing bulk operations on Linden trees.

The list is infinite, really. I have some examples below, but really, probably 80% of the feature suggestion forum could be implemented by a third party developer.

I humbly (me? humble?) suggest that this be the #1 criticial prioritiy after scaling / bugs / rendering.


Examples (but not limited to):

- better inventory management
- avatar npcs
- better building tools
- Import-Export Inventory from SL {encrypted}
- obsencity filter
- better calling card management
- better event announcement management
- better cache management
- general extension of toolbase for architects / city planners / other users
- REQ: optional Landmarks menu in the main menu bar
- shared inventory for groups
- etc etc

The list is pretty infinite.

Here are a couple of things I would do. I'd also use the Find/Profile style interface to create a wiki in the system that would automatically bring up articles based on your location.

I'm thinking that like most other tech, porn will be first.

Then maybe...some other junk.

Maxx, you do realize how club owners could streamline the striptease clubs with a client API and some basic AI code, right? No more crazy newbies running to the Alphaville Hearald scream "sex slave," no more having to camp the welcome area for newbs, just free-flowing sex and L$...

Prostitution is the oldest profession. Professions are the foundation of an economy. Automation strengthens economies. A client API would allow people to automate the oldest profession and strengthen the economy.

That alone should encourage the Lindens to focus on an API.

I'd probebly concider implementing support for VR equipment such as head motion-trackers, a glove, and stereo-vision for HMDs.

Mouselook is almost there. And SL did start as Philip's project to have something to play with his VR equipment at first.

Linear device input layer. User downloads a program that allows them to map mice and/or joysticks and/or keyboard keys to axes and buttons. Scripters would use LDIL to allow people to control throttle, steering, weapons, and whatever else on vehicles, avatars, etc. The spec would be open so that developers could write clients for whatever OS they are running SL in.

Of course by the time we get an API, SL may have something like that natively, who knows...

that's really the point isn't it huns. the place is advertised for it's innovative idea of letting the users provide the content. it only makes sense for them to expand that idea to let the users provide the features as well, at least the ones that can be constructed using a good api.

why isn't this their top priority after bugs and scales? why aren't we hearing their rambling ideas about it at every turn. why do they dismiss our questions and approaches with yeah we'll get around to that one of these days?

it seems to be a gaping hole in the catechism. i'm not complaining mind. they can do whatever they like; they're the company. it just confuses me when people preach one thing and do another.

Personally, I'm implementing an LSL editor that can integrate with the current editor for things like saving/syntax checking, but has a text-entry field that's outside of the SL app (hence preventing the lagginess that results from having every character a texture in a 3D scene)

-Adam

I want a client API, so I can...

1. Have an in-world wiki of sorts. If clients could grab webpages and display their text, we could build location-based wikis. An integrated wiki system, run onn a 3rd party server, would be a lot more informative than the current rating system.

2. del.icio.us styles capabilities. Imagine being able to press a button in your client, and have it register your location with a blurb about what you were looking about while you were there. Might clear the forum up a bit, spread word of mouth, and be an interface to a web-based, syndicatable feed to get bloggers interested in what was going on in SL.

3. iTunes/WinAmp control. I just like the idea of not having to leave SL when I wanted to fiddle with iTunes. I like the idea that you could use SL as a platform for widgets, much like Mozilla, that would allow steps toward a fully, 3D interface.

4. Hyperlinking. If we could fiddle with the client, we could embed hyperlinks into SL, and make it a more integrated, Internet application.

5. Ease off on the asset server. Imagine, when you wanted to send someone a note that instead of having to access your inventory, send the object to someone else's inventory, and then have them retrieve it from their inventory; you could offload 66% of that work to AIM, Jabber, or Y!IM. Use their servers to pass globs of text, and have a client-side reader extract it. That would take a lot of strain off the asset servers.

6. The above could work for games, like poker, too. Instead of passing textures for every card, you had a poker plug-in for your client. Instead of tapping the asset server, it sent a message via a background IM connection, and popped up a client-based window. Never bothering the asset server.

A client API would afford us a world of opportunities, as well as relieve a good bit of stress off the SL servers -- a decent plug-in would remove the need for a lot of XML-RPC and emailing. I look forward to when we get one.

these are good ideas. it's no wonder that ll isn't paying attention to such a small handful of posters though. perhaps we should host a couple of events and see if there's any more interest inworld than there is in the forum.

do people not want to be able to write their own feature set into the client?

Khamon I dont disagree that api is something we need the problem is releasing the code to allow api interaction.

For instance Window code for the most part is open but has a fully functional user interface accommodation for 3rd party add ons. However, the code for SL is still "private" by introducing an api they would have to release some of the code to do this and support 3rd party add on creators with specific data update information to keep the add ons working.

This I think is a scary prospect for SL because if they release their code whats stopping some developers from Creating another world using Bryce as the building tools and Poser as the true avatar creation environments and Vue de espirit for the world generation. Then on top of all that making the client side of the system require all these programs inorder to play it. Then you have an elite of the elite world of SL2.

Thats what I am seeing at the moment. Not being negative mind you as I would play the other..LOL

Shadow

I think a brainstorming event is a good idea. /stamp! It would be good to be able to discuss stuff in person, and maybe people could bring demos of things to show, haha.


Yes, but even more I want to be able to use add-on features written by other developers. Can you imagine what kind of Ivory Tower tutorials Lumiere Noir could create if he could show how building works by driving the build tools for you via an API?

I do have some concerns though. As a Mac user, I'd want to ensure that Macs don't get marginalized by developers writing plug-ins that work only on Wintel PCs. I also worry about trojan horses - it would suck to have someone write a plug-in that transfers L$ to them when I use it, or sends them my password, etc.

shadow, there's nothing stopping people from doing that now. in fact, i know people that are hacking around with it. i'm looking into it with a group of educational software programmers. the osmp project is working hard at it every day.

this is exactly the point. if ll want sl to remain the premier virtual world environment, they're going to have to provide the tools for people to cutomize the client to their own uses rather than have those people develop their own proprietary systems. it's getting to be all too easy to do, at least for windows, with things such as the directx and the .net sdks coupled with a good database engine.

they're going to have to stay a step ahead in order to not fall behind.

I agree khamon and Im glad you were able to decipher my intent from that simple rambling..

As a Mac user, I'd want to ensure that Macs don't get marginalized by developers writing plug-ins that work only on Wintel PCs.

I keep touting Mozilla as an example of a platform with an API, but that's because I would hope the LL people follow their lead. Make the API fairly simple: give us access to the UI, make sure the code that accepts UDP packets and talks to the OpenGL libraries is embedable, and/or give us access to sockets/IPC's so we can have our clients talk to other programs. Any two of those seem like all we'd need, and it seems like they're all fairly easy to implement on most any platform.

trojans and such things are a valid concern. i can see the mcaffee shield shudder every time i post about this.

olmy, mac users, being such a very small percentage of the computing public, are marginalized by definition. i use linux on several machines. i know, for a fact, that when really cool features hit windowsworld, and i want to have one too, i'm gonna have to either port it myself, or wait for somebody else to do it.

the same applies here. mac users, as a group, will have to reverse engineer your own plugs.

Thank you for starting this thread Khamon. I had no idea what the api really was, much less the benefits it could bring! So, just wanted to say thank you.

I just read the Avatar NPC thread. Here's something else we could do with an API. We could setup NPC's hooked into a database on someone's computer, and said database could be filled with news and bits about Second Life. You could bypass the Find window, and have an NPC give someone a landmark for a place. Have the NPC give them a notecard with instructions.

Just a thought.