Category talk:Data Formats

13:00:14 < ShockLinux> I've been for a the approach of having all info in different formats: 13:00:20 < ShockLinux> xml file links to a wwb file (which is what we have now) 13:00:28 < ShockLinux> xml file links to data in other xml file (maybe, but not very usefull) 13:00:40 < ShockLinux> xml file links to data in an xml tag in the same file (we need this the most) xx:xx:xx < ShockLinux> xml file links to data in a db (internal/external) (usefull for big stuff, like placenames) 13:00:47 < ShockLinux> which could be for a trail/placename(s)/etc

13:18:30 < ShockLinux> we need an in-xml version of all external files if you ask me 13:18:33 < ShockLinux> placenames too 13:18:50 < ShockLinux> then you make everything inside an xml first 13:19:05 < ShockLinux> in case you want more performance you convert it to the original form 13:20:02 < ShockLinux> that would also leave you with just one converter: xml --> wwb (or anything else in case your dealing with                      icons/placenames/etc) 13:20:51 < ShockLinux> fact is that people will come with format .abc and want to convert it to wwb 13:21:01 < ShockLinux> which isn't really working right now :) 13:24:37 < ShockLinux> you'd have a .xyz --> xml converter 13:24:47 < ShockLinux> and a .abc --> xml converter 13:25:00 < ShockLinux> and one xml --> (native; ie wwb) converter 13:25:10 < ShockLinux> of which the last one is optional 13:23:56 < ShockLinux> let the choice of converting or not over to the dev of the addon

13:18:51 could move to binary XML packet at some point < easy to automate and sencse > 13:19:52 <+jessi-nlt> but if you have a tool that reads/writes all WW binary data files, you don't need the duplicate plain-text files 13:20:30 if you move to an optional binary XML format you can jsut use standard XML tools :-) 13:20:41 <+jessi-nlt> perfect then :) 13:22:31 use a compressed form of XML i.e a gzstream 13:23:13 note there isn't a 'requirement' for a tag for every entry 13:23:15 <+jessi-nlt> if we put the plain text versions in CVS, the compilation process could include running script that "compiles" the XML versions into binary files 13:24:25 i.e 30  (30 coord pairs)  is valid XML 13:25:29 this would make adding trails much easier 13:26:00 have an XML and a NativeBinary representation for objects 13:26:28 provide translators to go between binary and XML