Paul Gilmartin wrote:

> READY                                                                         
>  
>allocate dsn('sys1.maclib','sys1.modgen')                                      
> 
> IKJ56700A ENTER FILE NAME -                                                   
>  
>foo                                                                            
> 
> READY       allocate dsn('sys1.maclib')                                       
>  

Just that without any other keywords? I believe default is DISP=OLD if I got it 
right.


>Empirically, if I allocate a single data set, ALLOCATE generates a DDNAME; 

Hmmm, interesting, I usually use both DSN and FILE unless I want SYSOUT.


>if I allocate multiple data sets, I must supply a DDNAME.

Apparently according the doc, in such concatenation, all datasets must be 
already cataloged, etc. Also, I see if you allocate a Concatenation, you must 
also FREE that whole Concatenation. It is something about managing the DCB 
attributes for all those datasets.

To me it makes sense, because you can only use Concatenation as input, not 
output.

Then at Example 18 of your manual, I see this thing:

<quote>
You cannot directly add another data set to a concatenation. There are two ways 
to add another data set to a data set concatenation: 

1.Use the FREE command to deallocate or free the data sets in the 
concatenation. Then reallocate the entire concatenation, including the data set 
to be added, using the ALLOCATE command.

2.Specify the REUSE operand with the ALLOCATE command when you concatenate. The 
REUSE operand specifies the file name being allocated is to be freed and 
reallocated if it is currently in use.
<end-quote>

Could this quote perhaps affect your experiment?


>(Have I inferred the rule correctly?)  I find no mention of such a restriction 
>in TFM:

Neither me.


>Is there any plausible rationale for this distinction?

If you get it, please post it here. 

As John Gilmore said, this extension to handle concatenation is somewhat 
botched, but I believe it is how the original SVC 99 was written in the first 
place. Just read up on SVC 99 and concatenation if you want to be confused... 
ouch...

>I have a credible, though illogical and farfetched alternative explanation:

Not farfetched.

>One might usefully create a single data set with DISP(NEW) (or MOD) with no 
>intent to use it immediately. 

Indeed, sometimes you ALLOC as per standard flow diagram, read your input and 
then found nothing to write at all.

Groete / Greetings
Elardus Engelbrecht

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to