You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@daffodil.apache.org by "Mike Beckerle (Jira)" <ji...@apache.org> on 2020/07/30 17:09:00 UTC

[jira] [Commented] (DAFFODIL-1742) TIFF - need DFDL schema and approach to TIFF

    [ https://issues.apache.org/jira/browse/DAFFODIL-1742?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17168079#comment-17168079 ] 

Mike Beckerle commented on DAFFODIL-1742:
-----------------------------------------

A proposal to the DFDL workgroup for extensions of DFDL to enable pointer based and offset-based structures was put forth.

This was reviewed with an attempt to cast TIFF in terms of it. It was found to have some challenges still. So we still don't have a proposed fully-complete set of new features that would enable TIFF.  BTW: even with those features, recursion was needed for TIFF, and there were still some other gaps. TIFF is effectively a linked list where each structure in the file has a pointer (file offset) that "points" to the next structure, which can be before or after it in the file. One issue was recognizing the end of the list. The other is that lists are generally recursive data structures. Lastly, if you can backward offset into the file, that probably allows parsing the same data over and over, so a parse may not terminate any more. DFDL's "forward progress" guarantees go out the window. 

Just a status report on where we stand on this as of July 2020. 

> TIFF - need DFDL schema and approach to TIFF
> --------------------------------------------
>
>                 Key: DAFFODIL-1742
>                 URL: https://issues.apache.org/jira/browse/DAFFODIL-1742
>             Project: Daffodil
>          Issue Type: New Feature
>          Components: DFDL Language, DFDL Schemas
>            Reporter: Mike Beckerle
>            Priority: Minor
>
> TIFF image format is not something DFDL v1.0 can describe.
> Nevertheless we need an approach that illustrates how to deal with it.
> I *think* this means a preprocessing program, and then a DFDL schema to turn the result of the preprocessor into an Infoset.  For unparsing, a post-processor also.
> The right approach remains to be seen, but we need to show how to deal with TIFF.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)