Page 178 - ITU Journal, ICT Discoveries, Volume 3, No. 1, June 2020 Special issue: The future of video and immersive media

ITU Journal: ICT Discoveries, Vol. 3(1), June 2020

5. CONCLUSIONS
The author, who has participated in the JPEG design and development effort throughout and who has
witnessed the incredible success of the JPEG format, has the following conclusions:
• The design concept to create a toolbox type of standard for JPEG was an excellent decision. The fact that
the common components were derived mostly from CCITT/ITU telecommunication and emerging PC
applications actually did not influence the wide applicability of the toolbox standard concept in
applications that had not been envisaged originally by the JPEG team. The selection of toolbox elements
in successful M-JPEG applications, which was definitely outside the scope of JPEG, is a good example of
that.
• The fact that some components, e.g. file format and colour model selection, were left out of the toolbox
by design was in the end not a limiting factor. Each application has found the right solution to fill those
gaps. Actually, for similar applications, ranging from the videotex photographic mode to the much more
successful worldwide web solution, to incorporate just the core JPEG components into a new protocol
environment was actually a useful feature.
• It is an irony of standardization that the majority of the originally targeted CCITT/ITU JPEG applications
in the end were not successful, but due to the flexibility of the toolbox principle new applications, e.g.
digital photography, web pages with photographic content and medical imaging, have played a role as
JPEG killer applications.
• Out of the JPEG toolkit, the definition of a JPEG baseline profile common to all applications and images
was a very good decision, and not only helped easier interoperability, but also provided a stable basis
for good quality images.
• High-quality photographic images on web pages proved to be one of the killer applications for the
success of JPEG. Nevertheless, the results of the development of a videotex photographic mode, as shown
in section 4.1.3, are very similar to this application, so could be easily adopted.

• It is fair to say that the first killer application was the inclusion of JPEG on the Internet and then in the
worldwide web. JPEG was included in web browsers from as early as February 1993. Thus, the boom in
the web was parallel to the popularity of JPEG. In the second half of the 1990s, another killer application
was digital photography, which got an additional killer application in the early 2000s, namely digital
photography by mobile and from 2007 smart phones.
• The patent policy of JPEG, which could only happen within an independent group with own rules and
procedures, was a very lucky choice: RF JPEG baseline with the possibility for some RAND options. That
JPEG patent policy then de facto disappeared after 1993 when the joint ITU and ISO/IEC rules for
collaborating work (ITU-T A.23 [5]) were approved. Then de jure a RAND patent policy was adopted for
all new standards, although JPEG informally often set targets to develop RF baseline standards.

• In general, JPEG derived benefit from the fact that until 1993 it operated according to its own rules in a
sort of vacuum, while its parents were much regulated SDOs, i.e. CCITT/ITU and ISO/IEC. That freedom
allowed effective, fast and innovative standardization work in JPEG, while the formal approval and
publication of standards by both SDOs after the completion of the JPEG specification took some time.

• The de facto collaboration of the IJG and JPEG was a lucky coincidence. The toolbox nature of the
standard and its RF policy were key requirements for the open source code implementation of JPEG by
the IJG. The IJG, after picking up the stable specification in 1990, could take the JPEG format as a baseline,
and combine it with JFIF for the missing components. The first code appeared in 1991, before the formal
approval of JPEG by CCITT/ITU in 1992. Later, it added further components, e.g. for progressive image
build up.

156 © International Telecommunication Union, 2020