Doc: IT/99-0109
Date: March 1, 1999
Project: Administrative
Ref. Doc:
Reply to: Monica Vago
Phone: 202-626-5739
email: mvago@itic.org
To: NCITS Members
Subject: AT&T Fast Track Proposal - NCITS Prima Facie Objections and AT&T Response
AT&T Response to Objections raised by Unisys, Xerox and IBM regarding the proposal for an "Image Compression Format for Digital Documents" (IT/98-0562)
2/19/99
AT&T RESPONSE TO UNISYS OBJECTIONS
(AT&T responses are shown in bold italics.)
Unisys is concerned about the submission defined in IT/98-0562, and would therefore like to register a prima facie objection to the processing of the document as a fast track standard. It is our belief that the introduction of the fast track process, while not limited as such, is intended to provide a process through which technology that had already received industry consensus in other fora could be brought into the formal process with a minimum of additional bureaucracy. It was not the primary goal to allow organizations - particularly single for-profit companies - to be able to bypass the technical consensus process already established. Thus the AT&T submission could set a disturbing precedent whereby NCITS could be used as a conduit to get ANSI approval (with minimal review) for specifications which have neither demonstrated consensus or proved technical merit. These concerns are similar to those expressed in the recent discussions on the JTC 1 PAS process and its use.
Thus far, Xerox, Hewlett-Packard and Sony have submitted proposals for consideration under the NCITS fast track process, i.e., there is precedent for companies to submit proposals.
AT&T does not agree with the Unisys position regarding a "disturbing precedent". The formal standards process is under continued criticism for being non-responsive and too slow in developing standards. The NCITS fast track procedure offers an alternative means for rapidly approving a standard through a consensus body, i.e., NCITS.
There are a number of questions which, when answered, would require thoughtful consideration before Unisys could vote on the document. Will the specification be turned over to NCITS L3 for maintenance/enhancement, and is there some assurance that AT&T would not further develop it independently?
As stated in the proposal (2.3), the "Proposed NCITS Technical Committee for Interpretations and Revision" is L3. After adoption as a standard, the appropriate technical committee identified by NCITS would be responsible for maintenance and enhancement.
Is there intent to further process this as a JTC 1 standard?
Not at this time.
What are the nature of the terms and conditions for licensing?
As indicated in the proposal (3.4.1), intellectual property held by AT&T would be licensed on reasonable terms and conditions and on a non-discriminatory basis, i.e., in compliance with the ANSI patent policy. Specific terms and conditions for licensing are typically not reviewed by NCITS.
These presumably would be answered in the submission, but we would be concerned that a reasonable analysis of the technology and the process involved might take more than the time available. (A perusal of the SD-2 indicates that approval for fast track can be either a 30-day letter ballot or a meeting vote. This would be followed by the public comment period, which can be as short as 30 days for an electronically available document. This means that (with the two week rule) the time from first sight to end of chance to comment could be as short as 45 days.)
AT&T is intending to comply with NCITS procedures. Despite the fact that the draft is an electronic document and therefore eligible for the shorter review period, AT&T is not opposed to the normal two month public review period if that would enable an organization to be better able to review the draft.
All of the issues above would apply to any submission that had not already had the benefit of widespread, if unofficial, support within the industry. In this case, however, there are specific objections to the use of the NCITS fast track method at this time.
1. The submission may conflict with work going on in NCITS L3, particularly the development of JPEG 2000. (Which we understand will be free of intellectual property considerations.)
The focus of early work on JPEG 2000 has been on coding of photographic images. The AT&T proposal is concerned with scanned document images, of which photographic images form only a small part.
AT&T maintains that its proposal is not duplicating JPEG 2000 since the intended use of the proposed standard is for documents on the web and not photographic images. AT&T notes that JPEG 2000 is at the Working Draft stage with a target date for CD in November 1999 and IS in 2001 (source: JTC 1/SC 29 web site, 2/16/99).
Regarding the comment "free of intellectual property considerations", AT&T notes that many NCITS standards include patented technology for which the patent holders agree to comply with the ANSI patent policy. In the proposal, AT&T states its intent to comply with the ANSI patent policy.
2. There could be "serious question as to its technical merit" because there has been no independent judgement of that merit. Some analysis of the specification is necessary to be able to say with any assurance that it is technically sound, and to date it has not been available.
AT&T is following NCITS procedures. According to the procedures, the specification will be made available following the prima facie objection period (during which the Unisys comment was made).
Both of these points indicate that a technical review and consideration of project overlap need to be undertaken, and they need to be undertaken by NCITS L3. This may involve some additional time; but there is no indication that the small extension required would cause significant detriment to the viability of the standard. We should use our fast track process, as intended, for cases where the "normal" process of NCITS would clearly be superfluous.
As stated earlier, AT&T believes that the fast track process should not be summarily discounted in a time when the formal standards process is under criticism for being too slow in producing standards.
Stephen P. Oksala
Director, Standards Management
Unisys
Malvern, PA 19355
AT&T RESPONSE TO XEROX OBJECTIONS
(AT&T responses are shown in bold italics.) .
Xerox raises prima facie objections to the fast track candidate standard entitled "Image Compression Format for Digital Documents (ICFDD)," IT/98-0561, on the basis that it conflicts with three NCITS projects. We enthusiastically welcome AT&T's participation in this extremely complex area of standardization and encourage AT&T to join us in two related NCITS committees. These are the NCITS L3 (Coding of Audio, Picture, Multimedia, and Hypermedia Information) and the NCITS L3.2 (Still Image Coding).
The fast track candidate standard, if approved for the NCITS program of work, stands to conflict with the following three NCITS projects:
(1) NCITS Project 1104-L, Lossy/Lossless Coding of Bi-Level Images
This NCITS project is the counterpart of JTC 1/SC 29 JBIG2. One of the components of the AT&T proposal is a coder for bi-level images. This component uses an algorithm similar to that being proposed for JBIG2. Because the scope of the proposal is less general than JBIG2 and intended for the web environment, the version of the algorithm in the proposal has fewer modes and options and consequently lower overhead. .
(2) NCITS Project 1106-L, Lossless Compression of Continuous-Tone Still Images
This NCITS project is the counterpart of JTC 1/SC 29 JPEG-LS. The proposal has essentially nothing in common with JPEG-LS which deals with lossless and near-lossless compression of photographic images. Where photographic images are contained in documents, ICFDD deals only with lossy compression.
(3) NCITS Project 1299-L, Joint Photographic Experts Group (JPEG) 2000 Image Coding System
The focus of early work on JPEG 2000 has been on coding of photographic images. The AT&T proposal is concerned with scanned document images, of which photographic images form only a small part.
AT&T maintains that its proposal is not duplicating JPEG 2000 since the intended use of the proposed standard is for scanned documents on the web and not photographic images in general. AT&T notes that JPEG 2000 is at the Working Draft stage with a target date for CD in November 1999 and IS in 2001 (source: JTC 1/SC 29 web site, 2/16/99).
Another related project is under consideration for discussion at the NCITS L3.2 meeting in January 1999, in Albuquerque, NM.
We have requested time on the L3.2 agenda to make a presentation on "mixed raster content (MRC)" compression for documents as part of the JPEG 2000 work. Can AT&T join the activity and help move it along?
AT&T has in the past attempted to further the use of MRC in JPEG 2000, in particular at the January 1998 NCITS L3.2 meeting. There has been no commitment by L3.2 or SC 29/WG 1 to include MRC technology in the standard. As indicated previously, the JPEG 2000 IS target date is 2001.
We encourage AT&T's participation in the related international standardization activities in the ISO/IEC JTC 1/SC29 (Coding of Audio, Picture, Multimedia and Hypermedia Information). The fast track candidate conflicts with three international projects in the ISO/IEC JTC 1/SC29 committee. These are: JTC1 Project 29.10, JTC1 Project 29.12, JTC1 Project 29.14.
AT&T has participated in the international standardization activities of JTC 1/SC 29 since its establishment (and prior to that as SC 2/WG 8-11). AT&T has provided editors for the original JBIG standard and initially for the emerging JBIG2 standard as well as other SC 29 standards. The AT&T proposal for image compression for digital documents on the web is not currently being standardized by SC 29.
The projects cited are JBIG2, JPEG-LS, and JPEG 2000, respectively. See above response.
Xerox lodges its prima facie objections to the fast track candidate, according to the NCITS / SD2, "Organization, Rules, and Procedures of NCITS (August 1998)," section 5.10, NCITS Fast Track Process. These objections basically encompass, and consider, the related standards activities identified in the NCITS document IT/98-0561.
Jean Baronas
NCITS Principal
Xerox
AT&T RESPONSE TO IBM OBJECTIONS
(AT&T responses are shown in bold italics.)
IBM has prima facie objections to the AT&T fast track candidate Image Compression Format for Digital Documents. The combination of JBIG2 (arithmetic coding subset), JPEG2000 (wavelets subset) and ITU-T T.44 conflicts with the AT&T fast track proposal. NCITS L3.2 is actively supporting the above emerging international standards and ICFDD contains duplication and conflict with existing projects.
The submission is targeted at a specific subset of the image compression arena, namely compression and presentation of scanned document images using the World Wide Web. In a sense, the AT&T proposal could be thought of as a containing similar functionality to parts of JBIG2, JPEG 2000, and T.44. However, the proposal includes technology specific to its targeted area, and hence is more efficient in that area than a profile built up from the emerging standards.
In addition, the proposal is ready now. The suggested combination of existing and emerging standards (particularly JPEG 2000) will not be available, according to JTC 1/SC 29 published target dates, until 2001.
In section 4 of AT&T's proposal "Related Standards Activities", the document lists JPEG-2000, JBIG-2 and T.44 and attempts to dispel any concerns regarding duplication by listing why the three standards individually differ from ICFFD. This approach does not address the intended use of these 3 standards in combination with each other as building blocks or toolkits for readily defining application standards which share common parts.
T.44 is intended to allow multi-level documents, but to do so it requires a rather heavy weight control structure, unlike the light weight IFF file format used in the AT&T proposal.
JBIG2, when adopted, will contain a mode in which the JBIG2 bi-level coding can be embedded in other compression systems. However, this mode is not as efficient as the bi-level coding in the proposal for document image compression.
The extent to which JPEG 2000 may or may not be used as a building block or toolkit is unclear due to the current development status (WD) of JPEG 2000.
We have additional technical issues that we can make available to any and all parties that may be interested.
AT&T believes that these technical issues should be addressed based on the specification that is submitted in the balloting process.
Best regards,
Ron
Ronald F. Silletti
Program Director of Standards
Intellectual Property & Licensing
IBM Corporation
North Castle Drive
Armonk, New York 10504