🚀 go-pugleaf

Sender: Bob Peterson <peterson@gemgrp.enet.dec.com>

The material in this report is copyrighted 1993 by Digital Equipment
Corporation.  Nevertheless any opinions expressed or mistakes in reportage are
strictly my own and not Digital's fault, as this paper has not been reviewed
within the company.  Whatta deal for you.  \bp



I attended the Second Annual Symposium on Software Globalization  Wednesday, 25
September.  It was held all day at the Hynes Boston in conjunction with the
Software Development Conference and Exhibition.  The Symposium was its own
micro conference piggy-backed on the administrative structure of the larger
conference.  This is a subset of the report I filed with my department.  I
thought I'd plug the conference a bit and contribute back to the people of this
list.  My editorial asides are in square brackets, thus: [ ].

I thought the conference was extremely well worth my attending.  It managed to
mention larger issues that I hadn't anticipated, and prioritized matters I had
been aware of before.  It was not particularly technical, though in the future
when they expand to two days this may change some.

I would recommend future conferences to project leaders or architects who need
to know strategies and processes for globalization.  [...] Quite definitely
anyone involved in
GUI-based products should send the P/L or architect.  The other person who
should go are managers concerned with marketing issues (and justifications)
around globalization.

I specifically requested more case histories and cultural studies, as well as a
deeper technical track for the future.  Such things could induce me to come
again.

[...]

The day was broken into two halves, a single track in the morning and a dual
track inthe afternoon.  At 4:50 we joined tracks to a master Q&A session.
The day was a series of presentations by individuals from companies with
experience in globalization or localization.  Globalization seems to be the
current politically correct term for internationalization (I18N).

Here's my detailed talk breakdown, by originally scheduled hours.  If the
conference had a serious down side it was all speaker who went on too long (and
thus cut Q&A time short).  A few speakers needed a better sense of value and
passion to their talks.


8-9 Why Globalize -- And How to Make It Happen
Kevin Cavanaugh, Director of International Product, Lotus

This was the most dynamic and interesting talk.  He defined the nature of
global business with examples of baseball (several countries play it but there
is no international league) and soccer/football.  He profiled a successful
copany globalization effort as a combination of company-wide mindset from
product inception, consistency and national diversity of the players.  He made
the case for globalizing (if we don't someone overseas will, and sell in the US
as well), and debunked myths about American leadership and American customer
being more demanding.  The nicest myth debunked by him and a few others was the
idea that simultaneous release was not vital.  The problem here is non-American
buyers will hold off purchase decisions if they see a release in the U.S.,
until the product comes out in their own language.  Meanwhile we lose profit
and possibly lose customers (impatience, competition).  He spent time
discussing how to implement a globalization process, which Digital has already
solved for the higher levels of the company.  Finally he mentioned many
benefits of globalization beyond a wider customer base.  For one thing a
globally marketed product, accessible to more people because they can
understand the language used, can add features.  His example was the addition
of "radar charts" to their charting software, an idea common in Japan and
introduced to the world through Lotus' globalization single-code-based product.

9-9:45 Internationalization of Software -- An Overview of the Key Issues
Tony Viola, Group Marketing Manager [ISE?], Digital

Tony sits just a few offices down the hall from me.  He focussed on product
management and design issues, more technical than Cavanaugh's talk.  His key
points were the requirement of a team mentality in the corporation to enable
the process, use of standards to reduce costs, and avoidance of localization of
non-globalized software (All-In-1 being his case in point).

10-11 Software Localization -- The Critical Success Factors
Claudio Pinkus, President, IDOC

[Yes, hyphenated talk titles seemed to be a mandated style.]
Mr. Pinkus started his own software translation company in the early 80s with
South American versions of CP/M and has grown from there to a very big concern.
I saw a great entrepreneurial spirit in him.  He addressed us with the pitfalls
he has seen companies make.  For instance, most companies rank localization
priorities as Time, then Price, then Quality.  The pitfall here is that lousy
quality can ruin a reputation, making it worse on the company than if they
hadn't localized at all.  It may take a couple years to recover mindshare on a
bad job of localization if the product was shipped too soon.  He addressed the
apparent conflict this creates against simulataneous release with the assertion
software should be globally tested in parallel and not as an add-on stage
late in the f/t process.

11-12 Going International -- Management and Marketing
Alistair Aitchison, International Product Manager, Novell

Alistair described ways a company new to the scene might implement a corporate
structure for writing globalized software.  He touched on structure models,
training ("enabling") engineers, translation structure, tools, testing and
control, documentation issues, what languages, support, and so forth.  Key
points I found valuable were: maintenance of a corporate term dictionary to
enable consistent product verbiage (many speakers emphasized this),
and that documentation volume and form itself is a cultural matter.  For
instance, the Japanese cannot deal with VMS' huge doc set (no shelf space!) one
the one hand, and on the other online documentation assumes a workstation on
every desk (not always true in Japan where some share access to workstations).

We split into two tracks and I took the technical track.

1:30-2:30 Making Software Japan-Ready -- A Case Study
David Pollack, Director of Japan Market Development, AEA Japan

He described his company, American Electronics Associates (AEA).  It is a
consortium of electronics and software vendors who want to sell in Japan.  He
characterized the Japanese market in contrast with the world market using sales
figures (20% IT expenditures, PC sales were overwhelmingly by native vendors
like NEC, and workstation sales split U.S./Japan with DEC under "Others,
27.5%").  The Japanese are generally more insistent on high quality and will
gladly lag behind versions.  Japan (and Asian countries) present a harder time
of maintaining one source pool because of unique I/O methods requiring extra
software.  He studied Oracle's bumpy ride into Japan.  Currently PC character
sets are highly fragmented.  Apple has the lead here in the U.S., but hopes are
Microsoft Windows' Unicode will catch on.  For today software must support all
existing character sets.

2:30-3:30 Language Technology -- Reducing Time to Market and Localization Cost
Michael Anobile, Executive Director, LISA (Localization Industry Standards
Association)

LISA is another conglomeration of vendors, an young attempt to make sense of
the localization industry.  This industry, btw, is expected to grow by leaps
and bounds.  Demand for services will so far outstrip supply that technology
must be applied to meet the need.  He broke down the industry into 4 size
groups, each much smaller than the previous: independents (kitchen table work),
small agencies, full service bureaus, and large companies.  LISA facilitates
the exchange of ideas on language technology and process, supports standards,
and provides training and several areas.  He spent much time on the concept of
consistent terminology from product inception to delivery.  Reusable
documentation is critical.  He spent time on the theme others also said was
critical: think globally from the very start.  Machine assistance is pretty
much limited to translation aids.  Rough draft machine translation is not a
mature enough tecnology.  Using a single system for all text (from initial
creation to final typesetting) was a great boon but not widely practiced today.
[I may have misunderstood him on the details of this complex topic.]  He spent a
lot of time on software tools available now.  Of interest to TLG might be
Globalware Inc's XL8 SuperPro, a localization management tool.  I noted it for
later consideration.  His final summary point was "Terminology management!"

3:45-4:45 Successfully Using International Development Tools
Ronnie Johansen, Director of International Product Development, WordPerfect
Corporation

This was a project supervisor/leaders nuts and bolts talk.  I admit I didn't
find this was as relevant as it should have been, either through Mr. Johansen
not knowing the audience or not preparing a clear enough presentation.  I did
pick out useful material.  Engineering must have clear controls over sources,
localizers should have limited access to source though they need context data,
Version control software with some kind of status ranking of modules is
important [...].  Everyone sat up for
his example of "pidgin english", using multibyte characters in English words,
expanding the word length and testing the software limits somewhat.
English-speaking developers can still use the software and get useful debugging
*before* trying to integrate with localization services.  Examples of pidgin
English:   "'`  ^`'"
         Fiiiileeeee    (in DEC 8-bit MCS: Fiomllejhik [may not survive Internet
                                                        mail])
The word is odd but clearly "File".  Welcome to Finnish-sized words!  This kind
of transliteration can be handled by a word mangler (with strictly reproducible
output) so automated testing is possible.  Automated testing is not always
possible [maybe some other speaker said this] because one may need to look at
GUI aesthetics.  [Such GUI testing should be done by foreign-based testers who
can check for color, pictures, etc. -mentioned by someone else?]  Further
testing problems include single/multiple EXEs (what happens if a bug is found
only in the Japanese version just after shipping starts?)

There was a Q&A session, which I took notes from for anyone interested enough
to press me for more information.  My attempts to summarize these talks has left
off many other valuable points made.  You want to know more, attend the next
one!  The next conference is March 15+16 in San Jose (2 days!) and then in
October in Washington D.C..  The brochure says to call in the USA 415-905-2784
(U.S. Pacific time zone, GMT+9 I believe).

\bob peterson
 digital equipment corporation
 (I'll respond to any private email *late* as I have much other stuff to deal