Noob thoughts/opinions
Moderator: Forum Moderators
if I would be targetting your sentences I would first quote them and then put my lines after that... or I would use 'Asmodai: blah blah blah'. These sentences were targetted to attitude that healers or some of them should be pacifists.Asmodai wrote: Wow, I don't know how you could possibly get this from what I was saying. Talk about taking a statement and then going WAY, WAY out to an extreme. All I requested is that healer units gain some exp for healing. I don't know how you took that to mean I wanted to be able to complete entire scenarios without any unit doing any killing. I'm not even sure where you got the whole pacifist idea from as I certainly didn't mention it. I'm perfectly happy with healers getting exp the same way as everyone else, but I ALSO would like to seem them get some exp when they heal, even if it is a very small amount each time (I used as an example possibly giving 1 exp for every 6 heals.)
Chinese warrior monks, celtic druids, aztec priests, apache shamans for apache, navajo, comanche, siberian tribers etc - many of them possess healing powers/skills and yet I don't see them as pacifists. Some of them were more peaceful than others.
- Miyo
On use of XML, I think that it would now be alot of effort to convert over to using it, even if we did decide it was the best course to take.
WML now has a preprocessor, which XML doesn't provide. The only viable solution [1] would be to use pre-processed XML instead of genuine XML, which would already lose most of the advantages of using XML, since standard XML tools would not be easily applied to Wesnoth configuration files.
Also, WML allows attribute and element names to have spaces in them, something prohibited by XML. Use of this is made in several places, and converting them over probably wouldn't be trivial.
Further, the interface to WML is used extensively throughout the application. If we were to switch over to using XML, we would have to re-write alot of code to use the XML interface instead of the interface to WML.
What's more, the two C++ XML interfaces [2] I have experience with I have found to be completely disgusting. I spent a large amount of time writing wrappers to both these interfaces to make them actually usable, something which other employees at my company greatly thanked me for. Unfortunately, those wrappers are owned by the company I worked for.
Once we decided on an XML interface to use, we'd have to make sure it was available on all target platforms, and we'd probably have versioning issues to consider, especially since XML-related technologies are only now starting to mature.
Writing our own XML parser is not a viable option. XML is a complicated language, and we couldn't hope to support all of XML. We could support perhaps a subset, but then we would essentially still be using an XML-like language instead of XML itself, it'd just be more XML-like than it is currently.
I am willing to admit it could have been a mistake to not use XML in the first place -- although I am not fully convinced of this, I do very much enjoy having a pre-processor available in WML, and I think the syntax is better -- however even if it was a mistake, I don't necessarily think that switching over to XML now is the best choice.
Finally, it should be noted that I devised WML by myself, before I dreamed Wesnoth would be so successful or popular. If I had known so many people would be interested in its internals, I probably would have been more likely to choose XML because of how widespread and standard it is.
David
[1] In case anyone was wondering, removing the pre-processor isn't a viable solution.
[2] MSXML/SAX and MSXML/DOM
WML now has a preprocessor, which XML doesn't provide. The only viable solution [1] would be to use pre-processed XML instead of genuine XML, which would already lose most of the advantages of using XML, since standard XML tools would not be easily applied to Wesnoth configuration files.
Also, WML allows attribute and element names to have spaces in them, something prohibited by XML. Use of this is made in several places, and converting them over probably wouldn't be trivial.
Further, the interface to WML is used extensively throughout the application. If we were to switch over to using XML, we would have to re-write alot of code to use the XML interface instead of the interface to WML.
What's more, the two C++ XML interfaces [2] I have experience with I have found to be completely disgusting. I spent a large amount of time writing wrappers to both these interfaces to make them actually usable, something which other employees at my company greatly thanked me for. Unfortunately, those wrappers are owned by the company I worked for.
Once we decided on an XML interface to use, we'd have to make sure it was available on all target platforms, and we'd probably have versioning issues to consider, especially since XML-related technologies are only now starting to mature.
Writing our own XML parser is not a viable option. XML is a complicated language, and we couldn't hope to support all of XML. We could support perhaps a subset, but then we would essentially still be using an XML-like language instead of XML itself, it'd just be more XML-like than it is currently.
I am willing to admit it could have been a mistake to not use XML in the first place -- although I am not fully convinced of this, I do very much enjoy having a pre-processor available in WML, and I think the syntax is better -- however even if it was a mistake, I don't necessarily think that switching over to XML now is the best choice.
Finally, it should be noted that I devised WML by myself, before I dreamed Wesnoth would be so successful or popular. If I had known so many people would be interested in its internals, I probably would have been more likely to choose XML because of how widespread and standard it is.
David
[1] In case anyone was wondering, removing the pre-processor isn't a viable solution.
[2] MSXML/SAX and MSXML/DOM
-
Pushnell
I know I'm tuning in late, but would a Markov Chain really be necessary? ... The game already runs fairly slow on my old P-300 without statistical name generation. It would seem all we need is a list of elvish-sounding first names and last names, combined at random for unit name. (Ditto with other races as they appear, but that's a different thread 
).
If you need a list of names, I'm sure the forums would take care of that in about 3.62 nanoseconds. (In fact, you may need to add a sticky about "NO MORE NAMES!!!"
)
).If you need a list of names, I'm sure the forums would take care of that in about 3.62 nanoseconds. (In fact, you may need to add a sticky about "NO MORE NAMES!!!"
)I don't think that it would take a significant amount of time. It'd be easy to tell too: the name generator would be run every time a unit is recruited, so if there's a pause when a unit is recruited, it means the name generator is slowing things down.Pushnell wrote:I know I'm tuning in late, but would a Markov Chain really be necessary? ... The game already runs fairly slow on my old P-300 without statistical name generation.
In my view, there are three main things in the game that currently take a while:
(1) blitting graphics to the screen
(2) the AI thinking
(3) loading the initial data set into the game
I don't think any game logic, other than the AI, takes a significant amount of time.
Well you'd think so, wouldn't you? But this thread: http://wesnoth.whitevine.net/forum/phpB ... .php?t=343 has been going for a while now with no responses.Pushnell wrote: If you need a list of names, I'm sure the forums would take care of that in about 3.62 nanoseconds.
David