How does the GIFT team engage other groups to get them to think internationally?

by Michael S. Kaplan, published on 2005/02/15 02:47 -05:00, original URI: http://blogs.msdn.com/b/michkap/archive/2005/02/15/372858.aspx

I think it went very well, though it was slightly under-attended due to Valentine's Day (though maybe I am just not a draw like I used to be -- people may be getting wise to my presentation style!).

As a bonus, both Brian Grunkmeyer (BCL dev) and Kit George (BCL PM), who post on the BCLTeam's WebLog) were there. They showed up at the very beginning of the meeting before the presentation and were talking with people who were at the meeting, asking them questions about how they use the .NET Framework. Both of them are very engaging in public so it was great that they were there in the beginning. I think I may try doing that at future meetings, it was very cool.

Anyway, I went through a whole bunch of the cool features being added on the globalization front in the Whidbey release of .NET. At the end of the talk there were some great questions, some of the concepts of which I will try to integrate into future versions of the talk. One of the most interesting questions was not about the talk itself but was one that I paraphrased in the title of this post.

(I think it was inspired by how I talked about how the linguistic and technical gloablization expertise on the GIFT team is often leveraged by other teams throughout Microsoft)

My answer was i think a surprise to him (it certainly was to me!). It may not be an opinion that everyone on my team agrees with, but I think I am right here....

I said basically that the best way is to work with the test teams of the various groups throughout Microsoft.

The logic is simple (though the answer I gave last night was an abbreviated version of what I give here!)¹ --

PSS (Product Support Services) are not always consistently the best good route because although they are our best customer contacts and the last line of defense when everyone else messes up and we accidentally ship crap, they have to fight hard to be noticed sometimes and it is always an uphill battle for them to effect change except when they have customer reported bugs.

UA (User Assistance)/UE (User Eduation) a.k.a. the people who provide help content are not always consistently the best good route because although they are the second last line of defense when everyone else messes up and we accidentally ship crap, they have to fight even harder to get noticed sometime and it is always an uphill battle for them to affect change.

Development is not always consistently the best route because being the "international developer" is in too many cases synonymous with being the "new developer" and the expertise is often not in a position that regularly shifts owners. In many cases there are important exceptions to this rule and when such developers exist we engage them and assist them. But it is hard to try to plan a company-wide strategy around the exceptions no matter how cool they are.

Program Management is not always consistently the best route for roughly the same reasons -- though the exceptions to this are also important and we work hard to support these people when they make themselves known or when we find them. But again it is hard to try to plan a company-wide strategy around the exceptions no matter how cool they are.

Because TEST has the job of trying to break the product before it is released so that users do not have to find it broken after release. They are our first and best line of defense when developers and program management mess up. They are recognized more than anyone else at Microsoft for finding by both automation and manual testing bugs that are interesting and/or important to the success of a product. And since the folks on the GIFT team can be a rich source of finding important and strategic bugs in software by helping them accept the core international scenarios that affect their products, they pay attention.

Once we engage those testers (of they engage us) and they start finding issues or problems or bugs or design flaws, developers and program managers also become more actively engaged in wanting to understand parts of this area of their product that they may not have fully explored earlier. Before the bugs are reported! And both devs and PM can after that even start to work more with UA/UE and with PSS to make sure issues that they cannot fix yet or which may not support best international practices can be well documented to allow users to get the products to work well. The testers also often engage PSS to help find customers who may have hit problems previously to help further support the need to handle international scenarios well.

They are awesome as partners in getting teams to recognize the importance of international (and other) scenarios which other members of their team are not always paying attention.

1 - Note also that this does not apply to the GIFT team, as being entirely focused on international scenarios from core globalizaion to localizability to typography to rendering to localization tools, we all end up being focused on the right areas without having to force the issues by proving the broken nature of scenarios (though even on this team the testers often have the best angle to affect change based on finding regressions and problems!

I have also experimented in this post with only syndicating the excerpt.... just to see how it works out. It turned out to be easy (if you ignore the two false starts in the first few minutes).

But I am curious what this will do to the site and who looks at it....

On topic (almost), maybe you could get the MSN Search team to think internationally (that is, if they want broad adoption). Google has its pages localized in pretty much all languages on Earth and the localized search works fine, but MSN barely has a few languages (I know, it's just v1...)

Off-topic, syndicating only the excerpt is a good way of preserving bandwidth; on the other hand, I like to read posts in the RSS reader; many sites syndicate only an abstract, but it's very short (just the title, or maybe 2-3 summarizing phrases). I think it's best to do it that way, instead of copying the introductory sentences; I also think it's a hard thing to do, because you'll need a catchy abstract.

Regardless of what you want to syndicate, I remain subscribed anyway :) You're THE authority when it comes to i18n issues and your posts are a great way of getting a broader perspective. Keep up the great work! And thanks!

Actually, the MSN Search team is pretty high on our "list" since they already use us (and many other cool technologies like MS Search and the word breaker/stem stuff from them and of course their own algorithms out of MSR).

I don't use exclusively Google *or* MSN Search these days -- best results come from doing both with the same search criteria since they are clearly using different search criteria and I never know which will be better. But I would love to see them push that language envelope!

Thanks for the advice on the excerpt as abstract; if I do that more often I will actually try to have it do just that. It will give me good practice at being catchy! :-)

And especially thanks for the kind words!

I like your logic on this one. Typically the best way to get noticed is by making it clear that you're working in the best interest of the customer, and the test team are high up on most people's radar when it comes to that. If they say there's a problem, then people will take usually take notice.

I don't mind syndicating extracts in your posts since I usually open them in the browser anyway (since it seems to do a better job with non-1252 codepage characters). Yours is about the only blog where that's true, though :)