[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Article on JAXP 1.3 "Fast and Easy XML Processing"

• To: Dare Obasanjo <dareo@m...>
• Subject: Re: Article on JAXP 1.3 "Fast and Easy XML Processing"
• From: Bill de hÓra <bill.dehora@p...>
• Date: Thu, 17 Feb 2005 11:30:44 +0000
• Cc: Michael Kay <mike@s...>, Elliotte Harold <elharo@m...>, Neeraj.Bajaj@S..., Chris Burdess <dog@b...>, xml-dev@l...
• In-reply-to: <830178CE7378FC40BC6F1DDADCFDD1D1042DAD79@R...>
• References: <830178CE7378FC40BC6F1DDADCFDD1D1042DAD79@R...>
• User-agent: Mozilla Thunderbird 1.0 (Windows/20041206)

Dare Obasanjo wrote:
>>-----Original Message-----
>>From: Bill de hÓra [mailto:bill.dehora@p...] 
>>Sent: Wednesday, February 16, 2005 5:39 PM
>>To: Dare Obasanjo
>>Cc: Michael Kay; Elliotte Harold; Neeraj.Bajaj@S...; Chris 
>>Burdess; xml-dev@l...
>>Subject: Re:  Article on JAXP 1.3 "Fast and Easy XML 
>>Processing"
>>
>>
>>>So given the choice of methods that take useless interfaces (since 
>>>they have to cast to get anything useful) and APIs with lots of 
>>>overloaded methods that support a rich set of data sources, 
>>
>>why would 
>>
>>>you think the former is better?
>>
>>Because it'll be more elegant to add metadata onto Source 
>>objects than continue with parametric polymorphism. It's hard 
>>not to conclude a fundamental abstraction is missing when a 
>>class has that many overloaded calls.
> 
> 
> And it's hard to argue that Source is actually a useful abstraction.  

Touche :)

cheers
Bill

• References:
  • RE: Article on JAXP 1.3 "Fast and Easy XML Processing"
    • From: "Dare Obasanjo" <dareo@m...>

• Prev by Date: RE: sequence with elements with the same name is not allowed?
• Next by Date: What should TrAX look like? (Was: Re: Article on JAXP 1.3"Fast and Easy XML Processing")
• Previous by thread: RE: Article on JAXP 1.3 "Fast and Easy XML Processing"
• Next by thread: RE: sequence with elements with the same name is not allowed?
• Index(es):
  • Date
  • Thread