|What's a Good Spread of Test User Agents?|
I want to make WURFL work
| 2:04 am on Apr 18, 2007 (gmt 0)|
I test my mobile sites with the Chris Pederick User Agent Switcher and Firefox. It works extremely well.
What I want to know is what's a good spread of phones that exemplify a number of technologies?
It's no big deal when all I'm delivering are very simple pages in WML 1.0 and 2.0, but I'd like to be able to do things like deliver more style and read GPS from a phone.
I'm not sure if that is possible, but it would be a MAJOR coup if I could do that. I have a site that has a location proximity search that delivers results based on their distance from a zip code or map coordinate.
If I could access a phone's GPS, then the search could be customized to the location occupied by the caller.
I'm using WURFL [wurfl.sourceforge.net] to identify phones, so I can figure out the capabilities of each phone fairly easily.
What I can't do so easily is go out and buy a whole bunch of phones to test with. I can use things like the Nokia and OpenWave sims, but they are just scratching the surface; plus, they tend to stick with newer tech. I can switch their user agents as well, but it's a lot easier to just use Firefox.
Can anyone give me a list of "must work with" phones? I can switch in their user agents, and even write scripts to simulate their connections.
| 6:32 am on Apr 18, 2007 (gmt 0)|
|If I could access a phone's GPS |
Not possible for Verizon, and probably most other CDMA carriers. For Verizon, and other carriers that lock their users into Brew with no Java, the only possibility is to develop a Brew app and have it accepted by and sold through the carrier.
Keep in mind that most cell phones don't have a stand-alone implementation of GPS. Most (other than some PDA phones, and then I think only with external accessories) use A-GPS, which requires over-the-air assistance from the carrier. This puts control of the GPS data squarely in the carrier's court.
| 12:16 pm on Apr 18, 2007 (gmt 0)|
That's kinda what I thought.
There's real privacy concerns with feeding GPS data to Web sites, but the rhetoric out there is so whacky and extreme, or so vague, that I finally went and asked one of the WURFL [wurfl.sourceforge.net] guys about it, and he gave a pretty similar answer.
[edited by: jatar_k at 12:29 pm (utc) on April 18, 2007]
[edit reason] no blog links, thanks [/edit]
| 5:10 pm on Apr 18, 2007 (gmt 0)|
I'm less familiar with GSM, since I'm a CDMA user, and not a developer for either. There might be more potential with GSM carriers, and with the few CDMA carriers that have more open policies than Verizon.
At this point, you would probably have to develop or obtain a Java applet that the user would download and install, as I don't think standards have evolved for passing GPS data via HTTP headers, nor have they been implemented in popular browsers.
| 3:23 pm on Apr 19, 2007 (gmt 0)|
|test my mobile sites with the Chris Pederick User Agent Switcher and Firefox. It works extremely well. |
I also use the same process to validate that I am returning the correct markup to each device. One note on this process though: just because the page renders and looks OK in FF does not mean it will work on the phone. The FF rendering engine will handle many things (typically related to invalid markup) that will cause a phone browser to choke.
| 4:58 pm on Apr 19, 2007 (gmt 0)|
|One note on this process though: just because the page renders and looks OK in FF does not mean it will work on the phone. |
Yup. Found this out the hard way. I test with a RAZR, as that makes a good "lowest common denominator" phone.
It has a fairly conservative size limit on pages.
I use Validome [validome.org] to validate my markup, as it validates WML 1 and 2, as well as a bunch of other XML stuff (although I find <oXygen/> is handier for that).
Pages validate fine in Validome, but the phone croaks with a "Page Cannot Be Displayed" error. I've learned to adjust my XSLT stylesheets to break up big pages.