Tr : Fwd: Re: WARC spec clarification on transformed WARCs

[Clément Oury]

Hi Steve,

Thanks for your reply.

I agree on your proposals. However, do you consider metadata/"ARC-Filedesc-Date" and warcinfo/"WARC-Generation" as new headers fields, or do you want to put it in the block?

Personally, even if it is allowed in the standard, I would prefer not creating new headers when it's not absolutely necessary. There are already a lot of optional headers...

We are allowed to write what we want in the blocks of warcinfo and metadata records, so I would prefer writing information such as "ARC-Filedesc-Date" or "WARC-Generation" as first line of the block.

What do you think about it?

Best regards,
Clément


Gordon and Clement,

thanks for your thoughtful response.

your suggestions sound perfectly reasonable. i'll try
to restate them below so that you can confirm that we
have reached a consensus.

given the following WARC states, the following conditions
should apply:

1) original WARC
  warcinfo record should serve as ARC "filedesc" record,
  with optional WARC generation

2) migrated WARC (ARC->WARC)
  a) warcinfo record should serve as migration description,
     warcinfo/WARC-Date should be migrated WARC creation date
  b) metadata record should contain content of ARC "filedesc"
     record, metadata/WARC-Date should be migrated WARC creation
     date, ARC "filedesc" date should also be in this record,
     and possibly the WARC generation could be indicated here
  c) response records should have the same date as each
     corresponding ARC record

3) second-generation WARC (WARC->ARC->WARC)
  a) same conditions as (2), and
  b) warcinfo record should indicate WARC generation

i believe we would need to agree then on the form of the
fields for:

  2b) original ARC "filedesc" date in migrated WARC metadata
    record, e.g. metadata/"ARC-Filedesc-Date" with ISO8601 date.

  1,2b,3b) WARC generation specified in warcinfo record,
    e.g. warcinfo/"WARC-Generation" with integer value
    indicating; 0=original WARC, 1=migrated WARC,
    2=second-generation WARC, etc.

i'm not sure if "WARC-Generation" is necessary, but it seems
potentially useful.



thanks so much,
/st...@archive.org

[siznax]

Hi Clément,

i agree. we should not create new headers for this information.
the named fields i suggested should be listed in their respective
record blocks.

/st...@archive.org