Automation in G11n end to end process

What can be or should be automated in G11n end to end process? Here are some candidates:

1. File verification, both English source files and translated files can be verified or validated using some tools.

2. Leverage from previous translations, this can be done in software
workspace, using a message database or even a translation memory.

3. Translation process management or workflow, a tool to streamline, simplify or standardize the translation process.

4. Testing environment setup like machine reservation, OS installation, application installation, etc.

5. Test suites execution.

6. Reports or metrics generation.

7. Release engineering: workspace synchronization, package, patch generation and integration, product image download, etc.

...

Posted at 08:33AM Nov 02, 2006 by byuan in Globalization  |  Comments[0]

Motivation issue in G11n group - Job design

Last time I mentioned that some young people don't like to work on
localization and/or  testing, then how to design the job in globalization group to motivate those people?

The following are the five core job characteristics that are
particularly important to job designs. The higher a job scores on each
characteristic, the more it is considered to be enriched, and the more the people can be motivated:

• Skill variety - The
  degree to which a job includes a variety of different activities and
  involves the use of a number of different skills and talents.
  
• Task identity - The
  degree to which the job requires completion of a “whole” and
  identifiable piece of work, one that involves doing a job from
  beginning to end with a visible outcome.
  
• Task
  significance - The degree to which the job is important and involves a
  meaningful contribution to the organization or society in general.
• Autonomy - The degree to
  which the job gives the employee substantial freedom, independence, and
  discretion in scheduling the work and determining the procedures used
  in carrying it out.
  
• Job feedback - The degree
  to which carrying out the work activities provides direct and clear
  information to the employee regarding how well the job has been done
  

Posted at 08:22PM Nov 01, 2006 by byuan in Leadership  |  Comments[0]

RFC 4646 and RFC 4647 about Language tags and matching in Web applications

RFC 4646, in combination with RFC 4647, is to replace RFC 3066, which replaced RFC 1766.


Tags for Identifying Languages 
Matching of Language Tags

Posted at 09:15AM Nov 01, 2006 by byuan in Globalization  |  Comments[0]

How to do i18n testing

Just want to share two great articles posted to i18n-prog@yahoogroups.com
from Addison Phillips who is the Globalization Architect in Yahoo! Inc.


International Testing Basics  

Learn To Type Japanese and Other Languages

Posted at 08:59AM Nov 01, 2006 by byuan in Globalization  |  Comments[2]

Motivation issue in G11n group (2) - Achievement motivation

After reading the following article, I decided to not write anything
but its URL, great introduction about Achievement Motivation.

Posted at 03:50PM Oct 31, 2006 by byuan in Leadership  |  Comments[1]

Price Elasticity of Demand

Last time I talked about the law of demand, it tells us that consumers
will buy more of a product when its price declines and buy less when
its price increases. But how much more or less will they buy? In order
to increase revenue, whether you should increase or decrease the price
of a product? My answer was to increase price before I learned about
price elasticity, now my answer is it depends :-)



It depends on the responsiveness (or sensitivity) of consumers to a
price change, this is measured by a product's price elasticity of
demand:

For some products - for example, restaurant meals - consumers are
highly responsive to price changes, they may order other items or even
never come again if a restaurant increases the prices of some items.
The demand for such products is elastic.



For other products, for example, salt - consumers pay much less
attention to price changes, substantial price changes cause only small
changes in the amount purchased. The demand for such products is
inelastic.



So my answer would be to increase prices of the inelastic products to
increase revenue, but might think about to reduce the price of elastic
products to increase revenue by selling more products.

One example is manay stores have the buy one get one free on sale
during weekend or holiday season, you will see most of those iteams are
elastic products, the inelastic products won't be on sale most of the
time.



How to calculate the price elasticity of demand?

Ed = [(change in quantiy)/(sum of quantities)/2] / [(Change in price)/(sum of prices)/2]

For example, if  Q1 is 1, Q2 is 2, P1 is 8, P2 is 7, then Ed = [(2-1)/(1+2)/2] / [(8-7)/(8+7)/2] = 5

if Q1 is 7, Q 2 is 8, P1 is 2, P2 is 1, then Ed = 0.2



A product is elastic when its Ed is larger than 1, it's inelastic when
its Ed is less than 1, if Ed is 1, it means the price change won't
affect the demand.

Posted at 03:34PM Oct 31, 2006 by byuan in Economics  |  Comments[0]

Measurement and best practices of i18n and l10n testing (2)

To better measure the i18n and l10n testing efficiency or the i18n and
l10n quality, it's very critical to do the following things:

1. To use standard keywords for different kind of bugs, the keywords
should be in a easy to understand and remember convention, such as
i18n-external, i18n-internal, l10n-external, l10n-internal,
g11n-external, g11n-internal, there shouldn't have too many keywords
like language specific or team specific unless it's really necessary.

2. Don't use i18n/l10n specific subcategories unless it's really
necessary, especially the i18n subcategory since i18n problems should
be part of the core problems, not problems of an additional feature.

3. To follow standard definitions of the bug priority and severity, so that the nubmers of P1, P2 or P3 bugs do make sense.

4. To generate metrics frequently (automatically) to compare progress
or trends among products, releases, teams, etc. Then to leverage best
practices from those who do the best or find the root causes and
develop an improvement plan for those who do the worst.

Posted at 02:28PM Oct 30, 2006 by byuan in Globalization  |  Comments[0]

Make or buy g11n automation tools?

In order to improve efficiency, we need to use more and more automation
in the g11n end to end process, from translation workflow and
translation memory management to localization engineering, from release
engineering to localization testing, from project management to
business operation, etc.



Should we implement those automation tools internally or buy commercial
products? There sure have pros and cons in both solutions, but the
following are some pros of commercial products in the context of a
market system which rely on specialization (Campbell R. McConnell and
Stanley L. Brue, 2005, "Economics"):



The majority of consumers produce virtually none of the goods and
services they consume, and they consume little or nothing of  what
they produce. The worker who devotes 8 hours a day to installing
windows in Fords may own a Honda. Many farmers sell their milk to the
local dairy and then buy margarine at the local grocery store. Society
learned long ago that self-sufficiency breeds inefficiency.



Human specialization - called the division of labor - contributes to a society's output in several ways:

• Specialization makes use of differences in ability.
  Sepcialization enables individuals to take advantage of existing
  differences in their abilities and skills. If A is strong and swift and
  good at tracking animals, and B is weak and slow but patient, their
  distribution of talents can be most efficiently used if A hunts and B
  fishes.
  
  


• Specialization fosters learning by doing. Even if the abilities
  of A and B are identical, specialization may still be advantageous. By
  devoting all your time to a single task, you are more likely to develop
  the skills it requires and to devise improved techniques than you would
  by working at a number of different tasks. You learn to be a good
  hunter by going hunting every day.
  
  


• Specialization saves time. By devoting all your time to a single
  task, you avoid the loss of time incurred in shifting from one job to
  another.
  
  

So if someone is already providing a great tool with a reasonable
price, should we implement another similar or worse tool with limited
internal resources?



But buying commercial products might demotivate internal people who are
very interested in the same technology or tools, then it would be very
important to talk to those people to understand their wants, to
motivate them according to Herzgerg's two factor theory which I 
explained in the "Motivation issue in G11n group"

Posted at 09:10PM Oct 29, 2006 by byuan in Globalization  |  Comments[0]

What's Self-interest

Self-interest, in the market system, is the motivating force of all the
various econmic units as they express their free choices. Sel-interest
means that each economic unit tries to do what is best for itself:
Entrepreneurs try to maximize profit or minimize loss, property owners
try to get the highest price for the sale or rent of their resources,
workers try to maximize their utility (satisfaction) by finding jobs
that offer the best combination of wages, hours, fringe benefits, and
working conditions, consumers try to obtain the products they want at
the lowest possible price and apportion their expenditures to maximize
their utility.

The pursuit of self-interest is not the same as selfishness.
Self-interest involves maximizing some benefit, and it does not
preclude helping others: A stockholder may invest to receive maximum
corporate dividends and then donate a portion of them to the United Way
or give them to grandchildren, a worker may take a second job to help
pay college tuition for her or his children, an entrepreneur may make a
fortune and donate much of it to a charitable foundation.

Posted at 11:36AM Oct 29, 2006 by byuan in Economics  |  Comments[0]

What's the law of demand? And why it's true?

I was aware of the law of demand before learning Economics, but didn't know there are even some interesting explanations :-)



The law of demand: All else equal, as price falls, the quantity demanded rises, and as price rises, the quantity demanded falls.



Why there is such an inverse relationship between price and quantity demanded? Three explanations:

1. It's consitent with common sense. Price is an obstacle that deters
consumers from buying, the fact that business have "sales" is evidence
of their belief in the law of demand.

2. In any specific time period, each buyer of a product will derive
less satisfaction (or benefit, or utility) from each successive unit of
the product consumed. The second ipod will yield less satisfaction to
the consumer than the first, and the thrid still less than the second.
That is, consumption is subject to diminishing marginal utility, so
consumers will buy additional units only if the price of those units is
progressively reduced.

3. The law of demand can also be explained in terms of income and
substitution effects: A lower price increases the purchasing power of a
buyer's money income, enabling the buyer to purchase more of the
product than she or he could buy before.  At a lower price buyers
have the incentive to substitute what is now a less expensive product
for similar products that are now relatively more expensive.

For example, a delcine in the price of chicken will increase the
purchasing power of consumer incomes, enabling people to buy more
chicken (the income effect). At a lower price, chicken is relatively
more attractive and consumers tend to substitute it for pork, lamb,
beef, and fish (the substitution effect).

Posted at 10:58AM Oct 29, 2006 by byuan in Economics  |  Comments[1]

Motivation issue in G11n group

Many young people don't want to do testing because they think coding is
more challenging or interesting, and many people don't want to do
localization because they think localization is mainly about
translation, is very repetitive and also not technical, unfortunately,
both of these happen in Globalization group, most of our day to day
work is about localization testing, so, how to motivate people in such
a group?



According to Herzgerg's two factor theory, Hygiene factors are sources
of job dissatisfaction while Motivator factors are sources of job
satisfaction, the Hygiene factors or dissatisfaction factors are
associated with the job context or work setting; they relate more to
the environment in which people work than to the nature of the work
itself.  While the Motivator factors are associated with the job
content or work itself - if you want people to do a good job, you
should give people a good job to do.

Improving a hygiene factor, such as working conditions, will not make
people satisfied with their work; it will only prevent them from being
dissatisfied, a low salary makes people dissatisfied but that
paying them more does not necessarily satisfy or motivate them.



So let's take a look at the Motivator factors: Achievement,
recognition, work itself, responsibility, advancement, growth, as what
I said at the beginning, work itself in G11n group is not very
interesting to many people, so we have to think more about how to
improve the other factors ... (To be continued)

Posted at 08:59AM Oct 29, 2006 by byuan in Leadership  |  Comments[0]

How to achieve efficiency?

One of my goals is to improve the globalization effectiveness and
efficiency, today I have learned more about the efficiency from text
book Economics - Principles, Problems, and Policies (Campbell R.
McConnell, Stanley L. Brue, 2005):



To achieve efficiency, we must realize full production, which mean that
all resources must be used so that they can provide the maximum
possible results, full production implies two kinds of efficiency -
productive and allocative efficiency:

1. Productive efficiency is the production of a product or service in the least costly way.

2. In contrast, allocative efficiency is the least-cost production of
that particular mix of products or services most wanted by our
customers.



To improve productive efficiency, we can outsource or offshore some of
non technical tasks, or use some automation tools, or to leverage the
best practices that have been proven in another project or team.



To improve the allocative efficiency, we will need to prioritize our
projects or initiatives according to the requirements of our customers
or business, and allocate our limited resources to the most critical
activities.

Posted at 11:06PM Oct 28, 2006 by byuan in Leadership  |  Comments[0]

I18n and non I18n bug submission and resolving trends comparison

The following graphic shows the different trends of the i18n and non
i18n bug submission and resolving trends in one small product.

One problem is both i18n and non i18n bugs were filed starting from day
10, bug non i18n bugs were getting resolved starting from day 100,
while i18n bugs were getting resolved starting from day 200, ideally a
product team should resolve i18n bugs  earlier and more quickly so
that l10n can be started early and released at the same time with base
products.

The second problem is i18n bugs were found continuously till the end of
the release even when l10n testing should have already started or
finished, it means a lot of i18n bugs were found during l10n testing
and thus  translation  and l10n engineering/testing will all
have to be reworked many times which means extra time and
resources/cost will occur.

Posted at 02:19PM Oct 28, 2006 by byuan in Globalization  |  Comments[2]

Measurement and best practices of i18n and l10n testing (1)

To measure how well you are doing i18n and l10n testing, here are some of the metrics:



1. How many i18n bugs are found before and after l10n testing is
started. Ideally most of the i18n bugs should be found and resolved at
the same time with the other functional bugs, and very few i18n bugs
will be found when l10n testing is started. The later i18n bugs are
found, the more l10n rework might happen, and the less time to resolve
those i18n bugs and thus the worse the g11n quality the product will be.



2. How soon can i18n bugs or l10n bugs be resolved compared with non
i18n and l10n bugs. I18n and l10n bugs shouldn't take longer time to be
fixed than the other bugs.



3.  How many i18n test cases or suites are reused for l10n
testing? The best practice is to reuse all i18n test cases and suites
in l10n testing, the difference between i18n and l10n testing is the
messages will be pseudo l10n messages during i18n testing but will be
real translations in l10n testing.



4. What percent of i18n testing and l10n testing is automated, the more
automation in i18n testing, the more configurations/platforms can be
covered. The more automation in l10n testing, the more time can be
spent on manual checking of the translations in the context.



5. How much leverage among different languages, builds and platforms
during i18n testing and l10n testing, the best practice is define a
complete matrix to rotate among languages and platforms for different
builds.

Posted at 12:43PM Oct 26, 2006 by byuan in Globalization  |  Comments[0]

How to dynamically change the display language of a web application based on user preference.

See the i18n sample codes of Java EE  at http://java.sun.com/javaee/reference/code/

Posted at 09:31PM Oct 25, 2006 by byuan in Globalization  |  Comments[0]