The Designspace

OS X 10.4.11 update: issues so far

Other Mac OS X updates have gone more smoothly for me. Here are the problems I've noticed so far:

Safari 3 seems to have trouble with interpreting clicks, particularly from devices like the Kensington Trackball and the Wacom tablet. Unfortunately those are the only two devices I use. It will start out OK, then pages will stop responding to clicks. Text areas will show that they have been clicked on, but cannot be typed into. It almost seems as if it is misreading the mouse position, but I am not sure.

Safari also suffers from slowdowns - I imagine this has to do with Safari icon caches, so I'll have to figure out how to empty them again.

I can no longer see or mount afp shares. I can mount windows shares, but cannot see the files on them. Can no longer see or mount either firewire drive.

Update to the update! The issue with mounting afp shares turned out to be unrelated to 10.4.11. Safari's response time, particularly in form-fields remains slow.

Detect browser support for style properties

It is common to use object detection to discriminate between browser capabilities (See Quirksmode: Object Detection.

However, sometimes it is necessary to detect whether or not a browser understands a particular CSS style property, and it is not obvious how to do this.

You can't test for the existence of the property directly but you can test for the data type of the value of the property. If the data type is anything but "undefined", the property is understood by the browser.

The example below shows a couple of rows from an automatically generated table of items. Each row has a corresponding editable version which shows up when the "edit" button is clicked. The original, non-editable row is hidden at the same time the editable row is displayed. If "display='block" is used to display the TR, Safari will render it as a TD. To correct this, "display='table-row" must be used on browsers that understand it.

A non-editable row, with edit and delete buttons.


When edit button is clicked, the non-editable version is hidden, and the editable version appears.


If display="block" is used to display the editable row, Safari and Firefox will interpret the TR as a single cell - probably correctly from a standards point of view - so "display='table-row'" must be used for those browsers.

IE6 renders it as a TR, as intended, and does not understand the "display='table-row'" property, so "display='block'" must be used for IE6.



A simplified version of the HTML source of the 2 rows is shown below:

<table border="1"> 
  <tr id="resultTR1" bgcolor="#CCC"> 
<td>Data goes here. &nbsp; <a href="javascript:editMe(1);">Click to edit this row</a></td> 
</tr>

<tr id="editTR1" style="display:none" > 
<td><form name="form1"><input type="text">Data goes here.</input>  &nbsp; <a href="javascript:form1.submit();saveMe(1);">Click to save this row</a></form></td> 
</tr> 
</table>

which creates a table with two rows, one of which is hidden until clicked:

Data goes here.   Click to edit this row

Data goes here.   Click to show non-editable row

Fixing a dead iPod

My old first-generation iPod finally gave up the ghost a few months ago. It showed the Folder-with-a-questionmark icon, no matter what I did. I fully charged it, attempted to restore it, reset it numerous times, all to no avail.

Since I had just purchased an iPhone, I didn't really need it, so I didn't bother to send it in and have it replaced (again!) by Apple.

However, the other day I got around to searching for answers on whether there was a trick to fixing it. Sure enough, there is a great post by Matt Hickey that explains how to fix an iPod that won't boot. Apparently this is usually caused by a hard drive that stops spinning because the metal case has warped.

The first thing you need to do is find a screwdriver or other tool thin enough to loosen the glue and pry open the case. I tried a number of items that were lying around, but didn't make any progress - the case looked as if it had never been touched. The ipod had been face-down on the desk while I was attempting surgery. When I turned it over - surprise! It was back up and running. Apparently the jolts caused by my poking around with nail files and screwdrivers were enough to jar the hard drive back into action.

Take a look at the article if your iPod won't respond to the usual treatments.

Notes after three years of using AdSense

When I first added AdSense ads to this blog, I had no idea what to expect. After three years, I would have to say that the results have been mixed, but overall I've found that AdSense is worth the effort, and to my mind, actually adds to the usefulness of the site.
For the first few months, AdSense ads earned only a few cents a day, with occasional spikes up to one or two dollars per day. Daily earnings slowly increased to an average of about $1.50, with occasional spikes to five dollars. That was enough to cover the cost of all hosting fees for all my sites, with some left over for new software and subscriptions.

With recent improvements in the layout and structure of my blog, ad performance has increased along with readership of the blog. Earnings continue to rise very slowly, but will probably never get much beyond hosting-fee money.

Google does make it difficult to determine which ads pay the most, which ads actually got clicked on, etc. but it is not hard to determine which content pages receive the most views and clicks. Google usually puts elearning software ads on my site's home page, and most of the clicks come from that page these days.

As a result, I assume that a lot of people come looking for information about elearning, which in turn has led me to post more tutorials and templates relevant to that audience. I actually find the google ads to be useful, myself. I can't click on them, but I've learned about products I wouldn't have found otherwise from them.

It's hard to say precisely what causes increases in clicks. Visitor traffic has gone up somewhat over the last three years, and spikes definitely occur when an article is mentioned on a popular site, like digg or adobe exchange, but spikes in AdSense earnings are often unrelated to traffic increases. It just seems that sometimes very high-paying ads simply decide to appear on my site, probably because of luck or some mysterious google algorithm.

One negative thing that does occur with great regularity is that earnings will slowly ramp up to a higher level over several weeks in a row, then suddenly drop back to a few cents a day, even if the traffic on the site remains the same. Perhaps Google runs out of high-paying ads, or perhaps particular ad campaigns come to an end.

Even with these drops, the ads still average enough to pay for all my internet experiments and hosting. Within the AdSense program there are various types of ads to choose from: text, image, video, referrals, search. The only ones that I've found to be worth using on this site are text and image. I use the search feature because it adds value to the site, although it earns nothing at all. The video and referrals don't add any value and don't earn anything, so I removed them soon after first trying them.

Optimizing is a big topic on the Google AdSense forums, but beyond a certain point, the types of positioning recommended for higher earnings detract from the usefulness of the site, which I am not willing to do.

In contrast to my generally positive experience with AdSense, Amazon ads earn almost nothing. I keep them on the site mainly because I think it is useful to refer people to books on relevant topics to what they are searching for, whether or not they click the ads. I also like to advertise books that I've found interesting, "just because". Other than that, I am not putting any time into "optimizing" for Amazon ads.

Google's AdSense program is probably one of the few affiliate programs which do pay for the content publisher to include on their sites. Although AdSense has encouraged the proliferation of spam "made for AdSense" sites, I think it has created an incentive for people all over the world to post more useful information on the web.

Testing whether a frame has been loaded from the frameset page

Have you ever tried to use a javascript to modify the properties of a frame? Particularly a javascript ON the frameset page? If you have tried, you probably know that timing plays an important role. If the frame in question has not yet been rendered, you will get javascript errors such as "element 'X' has no properties" .

I've found that even if the function call is placed at the end of the frameset page, by the time it is reached, the frameset may not be fully loaded. I'm not sure, but I think that sometimes the frames don't fully show up as "loaded" until their source pages are also loaded. And even worse, Internet Explorer may not even SEE the javascript if it is placed at the end of a frameset page.

The solution I've found is to use setTimeout as suggested by Johan Godfried on irt.org:

function changeframeSrc(){ 

     if(frames[0]){//alert('yes');

	 frames[0].location='page03.htm';}

     else{setTimeout("changeframeSrc()",100);}

      }

changeframeSrc();