Re: [Yaml-core] My own thoughts on YPath (for what they're worth)

[Clark C. E. <cc...@cl...>]

On Wed, Sep 24, 2003 at 10:01:33AM +0100, Ged Byrne wrote:
| I hope you don't mind me sharing some thoughts on
| YPATH.  YPATH should not be Xpath 

Well, my goal is two fold:
  (a) do not use anything from XPath unless there are lots
      and lots of reasons
  (b) do not deviate from XPath unless there are a few
      good reasons.

YAML is very different from XML -- I am not blindly 
idolizing Xpath. ;)
 
| The XML phile is used to Xpath. If YPATH is trying to
| be XPath it will never succeed, Xpath is just too big.

Agreed.  However, given the YAML <-> XML transformation,
it would be nice if _simple_ path structures happened
to work both in a YPath and an XPath processor and
return "equivalent" results.

| The XML Sceptic has rejected XML has hype. They have
| no need to learn the ever expanding mass of Xfamily
| specs because they're scripting language already gives
| them the abitlity to throw their data about any which
| way. The XML sceptic likes YAML because it is quick,
| convenient and lighweight. For example, it uses their
| favourite scripting languages datastructures, so they
| didn't have the displeasure of learning a new DOM. 

Yep.

| puts "\n//  *1"
| p y
| 
| puts "\n/product  *24"
| p y['product']
| 
| puts "\n/product/ *25,*34"
| y['product'].each{|n| p n}
| 
| puts "\nproduct/*/sku/  *27,*36"
| ((y['product'].each{|n| n}).each{|n| p n['sku']})
| 
| puts "\n/address/!~{sta}/ *15,*17"
| y.each{|k, v| v['address'].each{|k, v| p v unless
| k=~/sta/} if v.class==Hash && v['address']}
| 
| Without any work, the necessary features are almost
| there. I'm sure Perl and Python are no different. 
| 
| The rather complex each statements used in the later
| example to implement selection and recursion could be
| cleaned up by extending [] to allow expresions like 
| 
| [*] ( all )
| [**] ( resursive all )
| [*{if 'sku'}]
| [**{unless /sta/}].

I'd love to hear more.

Best,

Clark