The Presidential Election is Random

I don’t usually write about politics, but politics is a big topic right now, is it not?

So here is Mike’s Wild Postulus on Presidential Elections in the US: The result is random.

Is it not lost on people that the race is about a 50/50 split, and that if we randomly just picked two people out of a hat to run as the only two candidates, we’d probably expect a 50/50 split between them? This has happened two elections in a row.

My theory is that with all the mis-information out there, common folk (including myself) really don’t know who would make a better President. Kerry says bush is bad because of X, Y, and Z, and Bush says that Kerry is bad because of X, Y, and Z. But when both say the opposite things, only one can be right, and how can most people know? Well, we end up using our gut and guessing. And guess what, at the end of the day, we come out about 50/50 – the same as if we just did a blind election.

Prior to stumbling across this statistical fact, I used to think we should be moving toward a popular vote and away from the Electoral College. However, now more than ever I think our forefathers had it right – common folk aren’t really ready to decide who should sit in the Oval Office. Rather, we should probably beefen up the Electoral College. The people should decide who the candidates are (via the Republican and Democratic nominations), and then the electoral college should decide who is a better lawmaker, commander-in-chief, and leader of the United States. Unfortunately, the citizens of America just don’t have the data to know whether Senator Kerry has the right attributes or President Bush does. Maybe people closer to the two – people that know them personally and have worked directly with them – would be better able to make an informed decision on this very important role.

Is this pessimistic? Maybe it is. But the election speaks for itself- its a 50/50 split across the US. Isn’t that an amazing coincidence? The people of America are collectively voting “we don’t care – the two are about the same”. Its pretty amazing that while each of us can argue so fervently for our position, on the whole, we all wash each other out.

Corporate Email Users, Email Compliance and Google’s Desktop Search

Now that Google’s product has been out a few days, people are starting to write about it. One concern that is coming up is that of security, which you can read about in many articles such as this one.

One topic not yet mentioned is that for corporate users, Google’s product may also have severe legal implications. Its one thing to keep cached copies of webpages. But, keeping cached copies of email, which can contain important proprietary information, is another story. And as it turns out, if you delete email in Outlook, Google’s product still keeps its stealthy little hands on its own copy anyway. This means that users that may *think* they deleted that email when they actually did not. Unfortunately, I think this is a clear sign that Google still “doesn’t get it” when it comes to enterprise users and their search needs.

In 2004, many companies are spending huge dollars on making sure that their company is in compliance with new information responsibility laws. Companies are more responsible than ever for making sure that information about their customers is protected and not leaked. If your employees install Google’s software, and receive a transient email containing confidential information, Google caches it in the background, and won’t let you delete it without going back and manually looking for it. Most users won’t have any idea this is happening, let alone know what to do about it. And once that information is cached in a hidden place, its could accidentally slip into the wrong hands.

Likewise, if you archive your email in Exchange, or have corporate retention policies to delete emails that are older than 90 or 180 days, Google pays no attention. It happily caches your private information with no regard for your information lifecycle.

Lastly, there is the complaint that many users have already pointed out, which is that once the information cached, anyone logging into the machine as another user can now also see the cached copies of that email. So, no more lending your machine to your office mate while you are out to lunch.

If I were in charge of an IT department, I would be very concerned about employees installing Google’s product. Its a lot of legal risk.

Lookout Review in CNET

Another Desktop Search Review came out this week, and this one is from CNET.

Lookout didn’t fare as well in this review, where it ranked 3rd in a field of 5 (6 if you count Google, but they didn’t really review Google.

The reviewer was reviewing “products that index your hard drive”, and cited a few Lookout weaknesses, including lack of searching outside of outlook, not having the files/attachments indexed by default, and not having a file preview pane (fair enough). I think given that Lookout is clearly focused more at email indexing, this is not too surprising.

The reviewer never contacted Microsoft or Lookoutsoft for comment.

But its interesting to see the individual reviewer’s own bias in these tests. (Of course, I’m probably just bummed because Lookout wasn’t first!)

For example, the reviewer’s test had a corpus of “129 emails”. How can anyone review an email search tool with a corpus of only 129 emails? Do you have only 129 emails in your mailbox? The last time I saw a mailbox that small was the morning I started at Microsoft. Before the end of the day, my mailbox was larger than that!! So for this reviewer, he was searching for files. OK. Thats his slant.

Another interesting thing is that he didn’t even mention how well you can manipulate results once you’ve opened them. This indicates to me that he doesn’t really use a search product himself. We can all do the cursory glance at a product, and see which features “look nice”, but when you actually use it on a day-to-day basis, which one is most useful? This author didn’t mention the fact that you can’t drag-and-drop items from search result lists, or that you can’t delete/move/copy/reply/reply-all/etc from several of these products.

Anyway, I think the review was pretty fair for who searches filesystem stuff best. I hope to see a review sometime soon where the reviewer really gets a realistic view of how email fits into a worker’s daily life. Its easy to superficially “search my hard drive”. But, I don’t think thats nearly as big of a problem as “where is that email that Bob sent me with the accounting figures?” I’d love to see someone rank these products based on the latter.

Google Desktop Search Review

News just in- Google has released its long awaited Desktop Search product today. I’ve been very eager to see this product, so its exciting that its finally here!!

Here are a few notes for Lookout fans.

What’s Cool about it:

Indexes your files (including word, excel, powerpoint, and text), your browser history, email (outlook or outlook express), chat logs (AOL).

Integrated with Web User Interface
The interface to Google’s Desktop Search is exactly the same as their existing interface at Further, search results from your desktop look identical to those on the web, and are integrated into your existing web queries. Much the same way that Google “news results” show at the top of the search result list, desktop search results also now show at the top of the list. (See screenshot)

Once again, Google did a great job creating a small and lightweight product. The download is less than 500KB, installed its about 1.5MB. Once installed, the indexer runs in the background, and is not noticible. The installation is quick and easy – almost no effort at all.

What’s Not Cool about it:

Web Integration
The web-integrated user interface falls short for me for email searching. For example, say you are an Outlook user, and you want to do a search. First, you need to switch from Outlook over to the web browser. Search results are displayed only 10 results at a time (like a web page), and can only be sorted by date or by relevance (can’t sort by sender, recipient, subject, type, etc). When you do find what you are looking for, you can open it, which shows you the mail item rendered in the browser as HTML. To really use the item, you now have to click again.

Also somewhat limiting is that search results are not very actionable. Once you’ve viewed the item in HTML, you can reply, reply all, forward, or view it in Outlook. You cannot move the item, copy it, or delete it.

One interesting good feature is the “View entire thread” action which surfaces when a message is part of a larger thread. Clicking this view, allows you to see a summary of all other messages with the same subject. Its a very primitive threading UI today, but hopefully they will expand on this in the future.

Summary: email users are left with disappointing navigation through search results, and a ‘two click’ requirement to get to an email.

Advanced Search is AWOL
Unfortunately, the product has no advanced searching capabilities. What if you want to search for “emails from Bob”, “subjects containing ‘deadline'”, or for messages with attachments? Wildcard searching is also not available.

Summary: Its surprising that Google, the King of Searching, missed out on this capability.

Can’t search contacts or calendar
Unfortunately, while they did integrate email searching, you still can’t search for calendar or contact entries in your mailbox.

How does it differ from Lookout?

Google and Lookout take a fundamentally different approach to how we find content.

Google’s belief is that the Web is the killer app. All users will start with Web Searching, and then they use their strength in that area to show a little bit of Desktop search. For this reason, the Google user interface is the same interface as their web interface. If you believe the web search interface is the right one for your email, then Google’s product is great. As you do your web searching, you’ll now have your searches augmented with desktop results as well.

Lookout’s belief is that Email is the killer app, and that email searching and web searching have fundamental differences. As such, Lookout integrates tightly into Outlook for easy access (just click on the toolbar inside outlook), allows email-specific searches (like searching for senders, recipients, etc), allows users to act on search results in email-centric ways (such as reply, forward, move, delete, print, etc), and allows users to sort results in email-centric ways (sort by sender, recipient, subject, folder, etc)


Good entry-level product for Google. Great web-integration. I believe this product will change the way users use the web, and I look forward to seeing future products! I’m disappointed that it didn’t do more to really search your desktop. Its too web centric.

The biggest strength for Google is that it is so lightweight. Because it seems to run without interfering at all, users will not mind running it in conjunction with other products. Even if Google doesn’t yet offer all the features that a user needs, users may be perfectly satisfied to run two products – leveraging Google’s web integration strengths and also strengths from competing products.

Spyware Popularity

I was visiting the alexa site today and looking at the top-10 most popular sites on the web. Its mostly about what you’d expect. You will find:

Looks pretty much as expected, but who the heck is OfferOptimizer?

Well, turns out they are the recipients of some of traffic from some of the most commonly found spyware. They collect information off your machine and use it for their own sinister purposes.

I guess they must be happy – 8th most commonly accessed website on the net. Sigh. How sad!

Strings in C++ in 2004.

I wonder how many people are dealing with Strings today. How many programmers are stuck because they have one library using a std::string, another using an ATL::CString, and a third using a MFC::CString? Or maybe you are trying to interface to a BSTR? how about a LPWSTR? or even a plain old char *? Maybe you’ve got your own whiz-bang String class that’s even better….

As software designers in C++, we’ve just unequivocally and totally failed. C++ was heralded as the object oriented language to bring us interoperability and code reuse. Say what? How can we even fantasize about being interoperable when we can’t even agree on how to format a string?

I think Java and C# have two major advantages. One is garbage collection, and the second is a standard definition of strings.

How is it that here in 2004, we can’t figure out how to put an array of characters together in a way that works in all languages, and is interoperable with other programs? This is absolutely ridiculous.

Of course this is all well known and has been for years. As a struggling programmer sent back to the dark ages of C++ after having been spoiled by managed code, it just gets me particularly grumpy.