named.conf file in xml

[Kevin Darcy]

Tom Jones wrote:

>Look, I'm no fan of Microsoft, and I understand the "easy human"  
>readable config files. My point about MS is that they are embracing  
>standards like XML to make it easier to write tools and GUI's for  
>their applications. This is seen as great asset, so no matter what or  
>how , all you have to do is generate the xml file based on the schema.
>
Well, regardless of what Microsoft or their supporters may think, BIND 
is not "their application". The content and meaning of a BIND 
configuration is defined by ISC, and presumably ISC sees no value 
proposition is making those configurations accessible to XML-based tools 
and/or GUIs, or, if there is value, not enough for them to expend time 
and effort to implement an XML rendering of BIND's configuration data. 
If one were to speculate on why there is little or no perceived value, 
one might note that the vast majority of BIND installations are on 
Unix/Linux/BSD servers, and the vast majority of software that reads 
and/or writes those configs has been written by Unix system programmers, 
not app programmers or Windows programmers. For better or worse, XML 
hasn't exactly seen much penetration in the Unix system programming 
environment. So it's quite likely that the audience of "programmers who 
would prefer to manipulate/maintain BIND configs via XML" is vanishingly 
small.

>All I know is that, the more I've started to use XML the easier it  
>gets for writing webservices to which can control a lot of my apps  
>and their xml config files. This also allows me to not to need to  
>learn yet another config syntax since there is no real standard for  
>config files.
>
But, to do a competent job, one still needs to learn the *meaning* and 
*substance* of those configuration parameters, and XML really doesn't 
help you with that. The syntax is almost irrelevant, compared to the 
steep learning curve of learning the nuances of the DNS protocol itself, 
including all of its extensions, and ISC's reference implementation of a 
nameserver/resolver that speaks that protocol, with its extensions.

- Kevin

>
>tom
>
>
>On Mar 13, 2006, at 1:19 PM, Kevin Darcy wrote:
>
>  
>
>>Bit of a non sequitur there IMHO. If your point is "humans  
>>shouldn't be
>>fiddling with text files directly anyway, the 'Microsoft way' is to  
>>have
>>some fancy GUI as the human/config interface", then if said GUI is in
>>place, what does it matter whether the underlying config is a plain  
>>text
>>file, an XML file, or a bunch of registry keys? Or, did you have some
>>other point, and if so, what is it?
>>
>>- Kevin
>>
>>Tom Jones wrote:
>>
>>    
>>
>>>IMHO, It's this kind of thought and denial which allows Microsoft to
>>>grow and become more widely used and accepted.
>>>
>>>tom
>>>
>>>
>>>On Mar 10, 2006, at 3:40 AM, tsar.peter at gmail.com wrote:
>>>
>>>
>>>
>>>      
>>>
>>>>"XML is a standard" ???   Well, in the small context of 2005 - 2007
>>>>maybe XML
>>>>might qualify as "observed in the wild".   But surviving into a
>>>>distand
>>>>future ?
>>>>Allow me to doubt.
>>>>
>>>>Text files on the other hand will always be readable by humans and
>>>>manipulated by
>>>>computers.  Don't forget that the most importent issue with any
>>>>configuration file
>>>>format is to be understandable by the human reader ( who has to
>>>>understand it)
>>>>
>>>>
>>>>Request not granted. :-)
>>>>
>>>>
>>>>
>>>>
>>>>        
>>>>
>>>
>>>
>>>
>>>
>>>
>>>      
>>>
>>
>>    
>>
>
>
>
>
>
>  
>