Hi Michiel,   
      
   On 05 Jan 13 14:26, Michiel van der Vlist wrote to Ward Dossche:   
     about: "Comment-entry":   
      
    BF>>> The check sum.   
      
    WD>> You fail to convince me ...   
      
    MvdV> As I side with Kees on this one, let me give it a try.   
      
   Kees has a point, but just a minor one IMHO...   
      
    WD>> If that comment line is inserted at the end for example and included   
    WD>> in the checksum, I don't see what the problem is.   
      
    MvdV> No matter where that line is place, it will be specific for the version   
    MvdV> in question.   
      
    MvdV> So a nodelist compiled with another version of makenl, will not produce   
    MvdV> an exact byte by byte copy of your original even if everything else is   
    MvdV> exactly the same. And so the checksum will be different.   
      
    MvdV> This makes it a lot harder for the testers to verify if a new version   
    MvdV> does exactly the same as the older ones. A quick way to see if two lists   
    MvdV> are the same is  by comparing the checksum.   
      
   It only makes it a _little_ bit harder to verify this. Just use a diff program.   
      
    MvdV> Especially now in the testing fase, it is important that testers can   
    MvdV> easily verify if new versions or versions compiled for another OS   
    MvdV> produce the same result as Ben Baker's original version.   
      
   It would be a good idea to make this new function an option that is turned of   
   by default. That way it's still compatible with Ben Baker's original version.   
   And testers can take the complete prolog including the version string from the   
   original nodelist and generate an identical nodelist when the option to add a   
   version string is turned off...   
      
    WD>> The checksum is valid at my level for the file that I produce with   
    WD>> the software that I use. The fact that someone else with a different   
    WD>> software not inserting such a comment-line gets a different checksum   
    WD>> is irrelevant.   
      
    MvdV> For the end user of the product "nodelist", it is irrelevant. For the   
    MvdV> tester of the product "makenl", it is not.   
      
   It is not irrelevant to the user, if he is interested in what software was   
   used to generate the nodelist.   
      
    WD>> We're discussing makenl_ng here, not NLMAINT or NLMAKE or whatever it   
    WD>> might be. The advantage of getting the comment-line is staring   
    WD>> everyone in the face who is involved in nodelist-production.   
      
    MvdV> Put the information in the (L)DESC line of the accompanying TIC file.   
      
   Then the information will be lost if the nodelist is archived somewhere   
   without the file description.   
      
   Wilfred.   
      
   --- FMail-W32-1.65-B20130103   
    * Origin:  point@work  (2:280/464.112)