Cart
[Home] [By Thread] [By Date] [Recent Entries]
It's merely a strawman though to suggest the level of simplicity which might make a MicroSchema language feasible with supposed 'validating' MicroXML parsers. Maybe a '@ref' attribute on element is still necessary. I can't yet imagine how an 'element' element would be defined as recursive without being global and referenced within itself. There might need to be attributes for cardinality added too. Plus I remember my own arguments for allowing datatypes to be defined as either ordered or unordered arrays to compete with JSON. In any case I guess a typical MicroXML parser would defer validation so there is less importance how validation is done. ---- Stephen D Green On 17 December 2010 15:22, Olivier Rossel <olivier.rossel@g...> wrote: > On Fri, Dec 17, 2010 at 4:17 PM, Stephen Green > <stephengreenubl@g...> wrote: >>> Element referencing is not in this example. >>> In my opinion, this is a CRUCIAL feature. >>> >> >> I left out element/@ref because I've heard so many complaints that it makes >> a schema harder to read. You have to use a graphical editor which can >> resolve the >> references and check for any broken ones, etc. Reading a larger schema with >> references means switching back and forth between definitions such that you fail >> to get a picture of the overall structure. Without the refs it is easy >> to picture the >> XML itself. (see examplotron and CAM, for example where the structure can be >> 'seen' from the schema). Plus, if it is hard for a human to read it >> might also be >> logically harder to write a parser and / or GUI editor. > > I see your point : simplicity over all. > Ok for me. > (We will add referencing in MiniXSD :) >
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] |
