Dennis Forbes on Software and Technology   Subscribe to RSS


About the Author
Dennis Forbes is a Toronto-based software architect. While focused primarily on the .NET and SQL Server worlds, Dennis frequently ventures outside of this comfort zone into game development and image processing. He has been published in several industry magazines, has been quoted in the Wall Street Journal and has been interviewed by NPR.

He is a vice president and lead software architect at an innovative New York City hedge fund back-office services firm.

Dennis has been working on solutions for the financial, telecommunications, and power generation markets for over 13 years.


Recent Entries


The Feed Bag
Jan 11 - Answer: No
Jan 11 - The Git DVCS

 
Thursday, September 22 2005

I've removed the Google Adsense ads (they might still appear in some historic entries because of the way Radio Userland updates content - unless I change something affecting the page it won't upstream for just a template change. NOTE: They also appear in the "greatest hits" static collection). I removed them because they're ugly and distracting*, and they offered such a marginal return. I also didn't like that they could be taken as promoting a bias, in a small way implying deference and submission to Google. 

You might ask "Well then why did you add them in the first place?" Good question, and thanks for asking! Let's just say that I don't have total faith in the Do No Evil creed that Google publicly espouses. I can't help but think that Google has a financial incentive to boost the search ranking of pages that host Adsense content (it's brilliant really - You go to Google and do your search, awash in Adsense, all to shuttle off to sites filled with Adsense. It's an Adsense world, baby!). I like these pages to have some search significance, so this concerned me. Add the fact that Google needs to quickly index pages hosting Adsense ads (to allow for contextually keyed ads), offering another possible advantage of hosting their ads. Alas, I'm going to trust the impartiality of Google's search algorithms...

* Isn't it remarkable how Google snuck in as the underdog in search, and then slowly started integrating text ads. "They're different," the masses cried. "They're unobtrusive and low bandwidth!" Yet here we are today and Google is now serving up loads of full-graphic ads, all views tracked by the Google Brain (the same one that knows what you search for, your email account if you use gmail, and so on), and yet the Google honeymoon continues. I think Google has achieved some enormous technical achievements, and some of their products are extraordinary (Google Maps is a fantastic use of existing technology, making the competition look like garbage), but I just don't buy into the mythology that Google is somehow exempt from the forces that drive every other corporation.

  .NET   Blogging   IT   Personal   Software Development   SQL 
Thursday, September 22 2005

Find has some serious usability problems in most applications, particularly those that deal with complex content.

The general usage goes something like this: You are looking for a particular piece of text in a document or a webpage, so you pull up the handy find dialog, type in the desired text or pattern, and hit go. The text is found, hopefully, so the document scrolls some content into view. In amongst the reams of content there is the text that you are seeking, graphically inverted to draw attention.

Of course in a lot of current content, with varying sized text and different backgrounds and context, the colour inversion is next to useless. You're left with nothing more than the hint that the desired text is somewhere on the current page (some apps, though not all, put the found text in the center of the screen, but often that isn't possible due to document bounds. Other apps don't even properly scroll the found text into the view window, so you have to scroll backwards and forwards a bit to see if it's there). We're living in a world of extraordinarily powerful desktop computers: Use some of that fat client power and highlight the find hits better! Putting dancing angels around it. Have clippy run out and jump atop the found text (I'm only partly kidding). Do something to avoid the braindead functionality we have today. And please don't scroll the found text so that it's right behind the modeless find dialog box.

Earlier EntriesLater Entries

Dennis Forbes