Martin, you should always be free to design your own implementation if you
desire.  I don't think the point at which the data is collected is at issue
here but more the data format.  What you are suggesting is still each tool
provider doing it on their own resulting in incompatible data being
maintained for each tool.  If there is a reduced data set that could be used
by multiple tools, then the user would not have to maintain the larger ADATA
data sets and the tools could read the common reduced data.  This would
result in one copy of the published data format instead of one copy per tool
with a proprietary format.

Of course, the provided data would have to contain the data desired by all
tools and perhaps that's what John is trying to prevent.  If he provides a
published data format or API then he may have to entertain requirements for
changes to what he provides and it could make future enhancements more
difficult than it would be for a private format.  There is always
trade-offs.  It would benefit the tool providers and users but possibly add
work for the data provider.

Chuck Arney
Arney Computer Systems

-----Original Message-----
From: IBM Mainframe Assembler List [mailto:[email protected]]
On Behalf Of Martin Truebner
Sent: Tuesday, November 06, 2012 9:18 AM
To: [email protected]
Subject: Re: ASMLANGX files

Chuck et al,

Here is how I see it:

I use the ADATA data and extract what I need and combine everything into a
structure suitable for just me at compile-time (ADATA-exit).

There is almost no extra processing at execution time of the program that
needs debugging.

The ADATA is the HLASM-data as in the ASMLANGX-files but packaged and
prepared for external users. If you are uncertain about a certain piece of
information, then you need everything....

My point is: Reduce the data at compile via an ADATA-exit and if a new
version needs more data- recompile it with the new version and offer reduced
functionality without recompile.

My opinion. YMMV

--
Martin

Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE more at
http://www.picapcpu.de

Reply via email to