Thread: Windower: The Future of Spellcast (3.0 Discussion)    

A common issue we discuss on IRC (read: argue about) is the future of spellcast. Currently, the code base works very well and is by far our most popular plugin. However, a lot of the functionality is tacked on in a very non-optimal way.

We're allowing a lot of very powerful logic in the XML, but it can make the XML rather unreadable and messy. There have been a few ideas proposed and I'd like to see the community discuss them. None of them are going to happen, and as of now #1 is probably the 80% favorite. But you guys can change that, if you want.


#1: Spellcast remains the same. We clean up the code to be more optimal/efficient and your XML files remain the same with no changes. Your XML's will load into Spellcast 3.0 with minimal to no changes necessary. Very little new functionality would be added.

#2: Spellcast moves to a 2 part system. Your gear sets remain XML based for readability, but the rules/variables section becomes script driven (lua/python/javascript). For example, you'd be able to do something like:

Code:

if (FFXI.GetCurrentWeather == "Fire" && FFXI>GetCurrentTime() > 6)
{
      LoadGearSet("fire_weather");
}

This allows tons of extensions, efficiency and new things to be added. Plus, one limitation of spellcast is that you can only have one of each tag per rule (You can't do BuffActive="buff one" Mode="and" BuffActive="buff two" efficiently). This would allow total customizations.

#3: A complete overhaul of the system. XML would be gone completely, and the community would give the windower team complete control to rewrite the system to be super efficient, fast, responsive, modern and awesome. This would likely fall to Aikar who wants to rewrite the entire system using JSON and Node.js (for you nerds).

I'd like to hear your opinions, because I know people usually HATE change. However, without change, the system will not evolve. If you accept change, who knows what we could do the spellcast.

I'd like to hear your thoughts on it.

2 or 3, most fucking DEFINITELY.

Just because there's a newer spellcast doesn't mean all the losers who hate change have to change! :D

oh god, I just noticed this is a new thread, and not just the old windower future discussion thread. I should probably say something more constructive, I guess.

Spellcast as it is right now is pretty ugly. XML is fine for making gear sets, but to create a language like spellcast out of it is just a mess, and is pretty much the biggest problem that people have with using spellcast, imo. A new and cleaner scripting driven style would be a lot cleaner, and would also be easier for people to use.

Yeah, there's people who hate change. They can keep using their old spellcast if they want. New version would help all of the new/current people that don't use spellcast because of its complexity, and the people who want the change.

edit2: And personally, I have been wanting to redo all of my spellcasts for the longest time, but I've been keeping it off because it's such a hassle. I have been using stolen XMLs this entire time because I really can't be bothered to make my own out of scratch. And let me tell you, some of the people who made those xmls that used to be listed in the windower wiki (still are I guess? idk) had no idea what they were doing, cause I remember quite a few of them were just plain broken and I had to fix them myself, lol. I'm sure there are many people in a similar situation.

3, with emphasis on being new user/code stupid friendly.

#3, imo. I like spellcast as it is, but to have something faster, more stable, more customizable... fuck yea. If it stays as it is, though, I'll still use it and keep rewriting my xmls over and over.

Originally Posted by Alolbont

Rewrite it from the ground up, as long as its better no one will complain.

These are people getting something for free. They'll still complain.

Originally Posted by Septimus

And I doubt that you need it, but if you guys need anything from me (mainly game data-wise) feel free to ask.

I also doubt that it's needed, but same.

Option #3 seems optimal, but I love Python, so I'll say that #2 is a close second personally. Very close.

If you are going to take the time to improve it, do it right.



Edit: Actually, Python based development stuff would really interest me. If people need help with stuff or a tester, shoot me a PM.

Edit 2: Also, assuming you used Python, why would you need to keep the Gear Set information in an XML file? That just seems like an unnecessary data loading step. Why not just use dictionaries?

Code:

FireGear = {'Head': "Hat Name Here", 
            'Body': "Body Name Here", 
            'Neck': "Neck Name Here",
            'Ring1': "Ring1 Name Here",
            'Ring2': "Ring2 Name Here",
            }
            
            etc.

Spellcast is shitty and cumbersome; my one button windower macros are more versatile.
#3 and I might start using it again.

Fuck python, use ruby!

Originally Posted by arus2001

3, with emphasis on being new user/code stupid friendly.

^ This so much.

Code:

<if 3="Easy language">
	<pick="3" />
</if>
<else>
	<pick="1" />
</else>

I'm down for (3) if you guys implement the following:
1. A way to convert old spellcasts
2. A guide for writing using the new language
3. The ability to use include files

Assuming this arrives before I quit, (2) will be very important since I have zero programming training and anything I do know dates back to XML.

#2: Spellcast moves to a 2 part system. Your gear sets remain XML based for readability, but the rules/variables section becomes script driven (lua/python/javascript)

Can't get even more excited than this.

I've tried to avoid XML after my database course, but then spellcast told me to use it
Now I can be free again :D

ADD: That means I opt for both #2 and #3
ADD #2 : How much more efficient it can be do you think the custom scripting implementation compared to lua/phython/javascript option?