1. False Positives & Managing G11n in Sync with Development
Processes and technology have changed.
Must be executed by people
Human processes have concerns, efficiencies and if course conflicting situations
From Lingoport:
Adam Asnes
Michael Asnes
May 10, 2016

2. Agenda False positives background Static Analysis vs. Testing
Intro to Lingoport & Products
Continuous Globalization
Handling False Positives Demo
Code scanning, filtering and false positives management
What’s learned
Q&A, Follow-up Resources

3. False Positives
Test result which incorrectly indicates that a particular condition or attribute is present
Distracts from real issues
Can overwhelm results and usefulness
Source of friction when measuring a process for improvements
#1 Objection!!!
Coverity
VS. False Negative

4. Why False Positives Software is complex
When you are measuring complex conditions that in some cases might be a problem and others not, you should expect false positives results
The degree to which you can control and manage false positives will have a strong effect on measurement usability

5. Example False Positive
SimpleDateFormat formatter =new SimpleDateFormat("MM/dd/yy"); String dateString = formatter.format(new Date()); Using the default Java rule sets, this code will raise i18n issues. If the dateString variable is user facing, this code is indeed an issue to be fixed (for more information, check your rule set help pages). However, if the dateString variable is used for internal purposes, such as a in a support log file, it should not be modified and is a false positive.
Let's look at a couple of lines of Java code which creates a date string based on a hard coded date format using SimpleDateFormat:
SimpleDateFormat formatter =new SimpleDateFormat("MM/dd/yy");
String dateString = formatter.format(new Date());
Using the default Java rule sets, this code will raise i18n issues. If the dateString variable is user facing, this code is indeed an issue to be fixed (for more information, check your rule set help pages).
However, if the dateString variable is used for internal purposes, such as a in a support log file, it should not be modified and is a false positive.

6. False Positives vs Issues
Example False Positive:
Here, the string “message” is sent only to a log file. It won’t be user facing, and does not need to be internationalized.
if (!file.exists()) {
String message = "Cannot open resource. It's not a valid file or URL.";
Util.getLogger().log(Level.SEVERE, message); }
Example Issue:
Here the string “msg” is first sent to a log file, then displayed to the user. It will need to be externalized for translation.
catch (XMLStreamException e) {
String msg = "Error retrieving servers capabilities " + serverURI;
Util.getLogger().log(Level.SEVERE, msg, e);
controller.showErrorDialog(e, msg);

7. Static Analysis
Measuring conditions in source code, rather than having to actively test
Code scanning, looking for specific conditions:
Bugs
Security
I18n
L10n Changes to the resource files in the code repository(s)

8. Defect costs per stage When most i18n & L10n issues are addressed
Use static Analysis
For i18n & L10n
Traditional approach to SW Development in a well defined process

9. Static Analysis vs. Testing
Do both!
Testing requires that you hit all conditions to be measured
Testing is by its nature an iterative loop
Code, test, fix, verify
Can take more time, more to manage, more manual processes
More human error or omission

10. I18n Products and Services – Not a Localization company
Since 2001
Got started delivering i18n services with early forms of product
Products expanded and gained enterprise clientele

11. Lingoport Suite
Systems, automation & measurement to facilitate ongoing software internationalization and localization.
Globalyzer: Find and fix internationalization (i18n) issues that inhibit localization and global user experience during development. Catch up on i18n technical debt.
Dashboard: The hub of our suite. See it all, drill down, create notifications and manage the process. Bridge gaps between localization & development. Increase visibility and track globalization metrics.
LRM: Make localization proactive. Detect & manage changes to resource files in your source code. Streamline translation jobs. Automate localization updates between development and translation.

12. Lingoport Dashboard
G11n visibility over multiple products and projects
Drill down & Planning

13. Globalyzer Server: Customize and store Rules (no source access)
Workbench: Big i18n jobs, Configure rules, i18n focus
Lite: i18n check from developer IDE or automated check-in
Command Line: Automate i18n measurement from the repo
Globalyzer finds and helps fix i18n issues like:
Embedded strings
Concatenations
Locale sensitive methods/functions and classes
Programming patterns
Static files

14. Resource Manager See what’s new in resource bundles Automate Prep Kits
Automate file validation
Automate sending files for Localization
Via TMS or L10n Vendor Portal
Track it
Validate it when it comes back
Automatically insert it back in the repo if it passes
notifications as well as dashboard instrumentation
Automated Pseudo Localization

15. Continuous Globalization
i18n & L10n in Every Sprint and Release

16. Continuous G11n Some people will be focused on L10n some on i18n.
Both are important!
There is a validation measurement and process gate to each function
Nobody has to remember to run a script or perform something manual
Nothing is in excel
The logical flow is to make sure i18n is in line to support L10n
Translation can be very integrated:
i.e. client who integrates with WorldServer
Can also integrate with your L10n Vendor’s portals, other translation systems like LingoTEK
You can have L10n changes as small as one word, but they are automatically managed in and out of the repository
No developers need to be involved with L10n
Back in to build with file quality checks. Gives you more time for staging QA.

17. Continuous G11n in the Enterprise
Automation
Visibility
Metrics
Many small iterations among many teams. To support i18n and L10n systems support the churn.
It’s got to work without somebody remembering to make it work
Its elements have to be visible to enterprise constituents. If data lives on somebody’s hard drive, it’s dead to the organization
You’ve got to measure to improve across the enterprise
These concepts should not be new to software quality

18. Minimizing False Positives
Globalyzer
Scope (test directory, external library files?)
Filtering
Strings
Locale-sensitive methods, functions, classes
Programming Patterns
Static Files
Ignores
Commenting
In the dashboard

19. False Positive Filtering
Demo!

20. Rules vs Line Ignores

21. Lessons for the Enterprise
Your development teams are moving fast
Make i18n & L10n a measured & visible part of every sprint
Tune your processes to gain valuable measurements and insight, while cutting out false positives

22. Questions & Answers: Contact Resources
Adam Asnes Michael Asnes Lingoport Suite Extensive Services Training – training.lingoport.com
lingoport.com/blog
lingoport.com/resources
wiki.lingoport.com

23. Question #1
What programming languages have the most issues?

24. Question #2 How much will I have to filter?
I like LingoTek as they have many connectors to content managers, including AEM.
We of course connect to source code repositories, which are different.
Connecting from our system to translation systems and portals is pretty versatile and easy and described in our Wiki. Happy to discuss with you in detail outside the webinar.

25. Question #3
Do I have to have a person dedicated to rules? How do I manage upkeep?
Top 3 –
Clear business case expressed and understood across the company
clear and scalable i18n/l10n requirements & development process,
Connection to their global customers for feedback,
Yes, but they don’t always get a localization specialist. The L10n manager should be managing, not messing with files. Representing on some scrum teams or overseeing that L10n is part of the feature set for development and testing.