i’m a tabbed-browsing fanatic (most of my sessions have 20+ open tabs), so one thing struck me right away: new tabs don’t appear where they used to. in all prior versions of firefox since they added tabbed browsing, new tabs appeared at the end (far right) of the tab list. suddenly, in 3.6, they changed the default model to one in which the new tabs are inserted right next to the current tab. this is seriously flawed, but not for the reason you may think.

i don’t have an issue with the behavior itself — Google Chrome and IE 8.0 both have this behavior for their tabbed browsing. it doesn’t even deeply bother me that they changed the default behavior rather than rolling the feature out as an option in this release and then changing the default in a future one (a model i prefer). my chief complaint is that they took a core browser behavior, changed it without notice, and didn’t provide an option in the preferences for changing the behavior back!

i’m not the only person to complain about this, and luckily there is a fix. however, your average firefox user is not going to know that they can hack the about:config registry.

frankly, this is poor product management on Mozilla’s part — and a rare miss for a team that has historically done an excellent job of keeping their users in mind.

for my ‘day job‘ (oddly named since i often find myself working on it at night), i manage a website that is used by thousands of health care professionals around the world. in order to keep things running smoothly on the site, we have a series of tests that we run every time we make a change to the code powering the site. these tests let us know that the code we have just written (often for a new feature) hasn’t broken the expected behavior of existing functionality.

these tests are Very Important. i cannot stress this enough. if you are a software engineer, a software QA engineer, and especially if you are a support engineer, i am preaching to the choir.

i’m not saying that our test suite is perfect. but we have one, and we add to it regularly. we do this because the penalty for failure to test can impact thousands of users and in the worse case, drive them away from using our site.

i also run my own site. for many years, i wrote the code behind the site myself, but over time, i started to realize that i was essentially building from scratch what the folks over at Wordpress had already implemented quite well. since they test their own code, i had no incentive to create a test suite for my site or my wife’s blog.

about 6 months ago, i upgraded Wordpress from the absurdly old 2.0.x to a more modern 2.7.x — an upgrade which went seamlessly and also allowed me to use their new one-click upgrade feature (i.e. no more unpacking tarballs and editing config files). however, the feature was DOA.

after some digging, i found that i needed to move to MySQL 5 in order to take advantage of the quick-upgrade feature. the data migration wasn’t painless due to the vagaries of our hosting provider, but after some massaging, i got things ported over. next, i found that an upgrade from PHP 4 to PHP 5 was required — luckily, i just had to add a line to the .htaccess file in the root directory. with a few other minor tweaks, we were good to go.

earlier this month, my wife tried to log in to the administration panel of a custom CMS she purchased a few years back. the login failed, as did the contact form submissions she had been receiving for a few months. i looked into the issue and found portions of the CMS code were not PHP 5 compatible (despite 5 being theoretically backwards-compatible with 4). turns out that adding the .htaccess file to the root directory, as my hosting provider had recommended, caused PHP 5 to be used in all subdirectories.

why didn’t i catch this earlier? well, i was making changes i thought only impacted our Wordpress installations. i had not touched other databases or code installations. since my wife’s wedding website appeared to work, i did not think to test features like the CMS admin panel or contact form.

why was it a big deal? if the issues had just impacted our blogs, it would not have been. however, the contact form on my wife’s wedding photography website is used to field questions by prospective clients. empty emails mean missed opportunities. the good news is that she’s putting her business on hold while we adjust to becoming parents, but this sort of mistake is still relatively costly.

would testing have helped? certainly testing of some sort (manual) would have helped. automated testing would have been harder to set up, since we had a bunch of different interfaces, but something like Selenium could have done the trick.

why don’t we have automated tests for our personal websites? discuss.

first impressions:
* the interface is a wee bit messy. it looks a lot like an email interface (even uses the terms “inbox” and “spam”), but there are a number of controls in the list and detail frames. it also seems as if i can move each frame around… but i can’t. i can just minimize them.
* i like the fact that it automatically pulled my contacts from Gmail into the contacts list, and just those people who have Wave.
* i dislike the fact that i received new wave content over the past few months, but never received an email letting me know. i realize that Wave is supposed to act as a sort of replacement for email, but why am i going to check it if there’s nothing yet going on? talk up your product a little more, Google.

my first wave:
* received a wave from my friend who invited me. it seems empty. is it weird that you can start an empty Wave?
* the playback option seems neat, but i’m not sure when it would be useful. perhaps from a mobile device…
* this wave doesn’t have a subject (or text, for that matter), so where i’d expect the subject to be, there’s a list of the avatars for the participants in the chat. pretty straightforward how to add other people to the discussion / wave.

my other waves:
* ok, now there’s some content. my friend sent me a wave that includes an embedded photo. he also references adding ‘robots’ to the wave, and it seems as it on this wave there is an ‘email notifications’ robot. doesn’t it seem silly that you have to add a robot to send you emails? shouldn’t that be one of the built-in options for Wave? couldn’t i just set that app-wide if i wanted? either that, or a follow/unfollow via email model would work.
* looks like they’re already using a follow/unfollow model to keep or remove things from your inbox. unfollow makes sense to me… it’s like ‘mute’ in Gmail. follow doesn’t make as much sense. the action exists in your inbox, and if something’s in your inbox, doesn’t that mean that you’re already following it?
* there’s a ‘draft’ checkbox on my reply, but it won’t let me check it. unfortunate.
* oh… i get it. playback shows the progression of the wave. when you make something look just like audio playback, with controls and everything, your users expect it to be audio. just saying.
* apparently, i can edit my friend’s post. when i do so, it says that that post is now by him and me, but otherwise, there’s no indication in the text about who wrote what. in that way, it acts a lot like a Wiki. interesting.
* almost constant saving of text means that you don’t lose what you’ve written. nice.

other:
* you control your Wave settings in a wave? i know that we’re all supposed to buy into the Wave model of the world, but can’t you just put settings up in the top right nav, like every other Google app?
* when i click on my avatar in a wave, i see my info: an email address and a website. apparently i have a wave (@googlewave.com) email address? that’s news to me — haven’t seen that listed anywhere else on the page. if i can email info to myself and then discuss it with others, isn’t that an important feature?
* i tried emailing my own google wave email address. after 5 minutes, still no new waves.
* why is so much real estate dedicated to the list of waves? i would prefer the Gmail model, where the list of waves is the default but is replaced by the content when you click on one. This 3-column approach is very Outlook.

contacts:
* my contacts list looks a lot like google chat — i can even set a status (by clicking on my avatar). but it’s not chat. if you click on a contact, you can start a new wave with them, find waves by them, and ‘ping’ them, which i assumed was like ‘poking’ someone in Facebook.
* the ‘ping’ is actually just a wave, but one that renders itself initially as a chat, a sort of inverted version of the Facebook chat.

praise:
* Wave is an interesting concept, sitting at the collaborative intersection of chat, email, wiki, and document editing.
* there are collaborative elements here that could be split off and re-purposed elsewhere, perhaps integrated into other websites or applications. from what the Wave team has said in the past, it appears an API of this nature is forthcoming.
* i think the real value in Wave comes from the real-time aspect — being able to very rapidly interact with your fellow collaborators.

criticism:
* let’s assume that Gmail let you embed objects (images, maps, etc) right into your emails. let’s also assume that editing someone else’s text is something you’d rather do in fully-featured app like Google Docs. what value is Wave adding? it’s a confusing email client (compared to Gmail), sub-par document editor (compared to Google Docs), and a middle-of-the-road wiki (compared to MediaWiki, PBWiki, et al).
* Wave is still pretty rough around the edges, and while Google has been good at generating hype (hour+ presentation, limiting invites, getting good press via major news outlets), they have been operating in a “here’s a tool — figure out what it means for you” model. not a terrible model, but certainly not the way you’re going to revolutionize email and overthrow the status quo.

—

update:

let me preface this update by letting you know that when it comes to software, i’m a notorious non-instruction-reader. i believe that most software should be intuitive, so i wrote my review before reading or watching any of the instructional information. after writing the initial review, i watched some videos from the Wave team and learned some things:
* apparently you can reply within a Wave ‘blip’ (creating a sub-’blip’) by double-clicking on the blip. pretty cool for in-line comments.
* there are some Wave use cases listed by the Wave team: organizing events (ok, but i’d rather have it tied into my calendar); meeting notes (i’m not a huge fan of extensive meeting notes, but this would work well for those who are); group reports and writing projects (as i mentioned before, i’d rather use Google Docs); brainstorming (yes, yes, yes!); photo sharing (meh. i’d rather use Picasa, Facebook albums, Flickr, etc)
* i sent that email in to my account a while back, and still don’t have a new wave. i even tried sending from another email account, but no dice.
* robots seem pretty powerful (Blogger, Twitter, email notifications, etc), but i can’t seem to see an easy way to add these robots to my contact list. the UI doesn’t make it clear, and the short tutorial videos all assume that you already have the robots in your contacts (which i don’t, in case that wasn’t clear).
* Rosy the robot does automatic translation, so people who speak different languages can communicate — ok, that’s just amazing.

why fixed-width fonts? well, for hackers like me, code and “ascii art” display properly in a fixed-width font. fonts like Arial and even Times use different widths for different charaacters, screwing up the original formatting of certain emails. for instance, this:


+ Col1 + Col2 +
---------------
| data | data |
| mmmm | mmmm |
| wwww | wwww |
| oooo | oooo |


ends up rendered like this:

+ Col1 + Col2 +
—————
| data | data |
| mmmm | mmmm |
| wwww | wwww |
| oooo | oooo |

recently, they added a feature through “Gmail Labs” that toggles monospace, but it has to be turned for each and every time you read an email. a global setting would have been much nicer.

so i went searching, and this is what i found:

1) if you use Firefox, just install the “Stylish” extension and add a style for it.
2) if you use Prism, go to the unpacked bundle of the gmail webapp (the one on the Prism site unpacks to a directory called “google.mail@developer.mozilla.org”). in there, create a file called webapp.css (you should see webapp.ini and webapp.js already). add the following to that file:

.geTjV, .ArwC7c, .iE5Yyc, .uQLZXb > textarea {
font-family: courier !important;
font-size: 10pt !important;
}


enjoy.

1. Gmail (personal)
2. Gmail (work)
3. Google Calendar
4. Google Docs
5. Google Groups
6. Facebook
7. LinkedIn
8. Google Reader
9. Remember The Milk
10. Joes Goals

for a long time, i used Firefox’s tabbed interface to manage multiple applications at once. on the plus side, all of my webapps were in one convenient location. unfortunately, this also meant that one application could dominate/crash the browser (i’m looking at you, google docs). i also like to alt-tab through my list of applications. the one flaw that really sent me searching for a solution was how difficult it was to log into two Gmail accounts at one time. the ietab plugin worked well while i was in Windows (ietab uses a separate set of cookies), but i’m running Kubuntu full-time at work, so that option became less practical.

eventually, i found an under-hyped offering by Mozilla Labs: Prism. web 2.0 applications have removed the need for many of the typical “web” features — link bars, navigation buttons, etc — and have started to look more and more like desktop applications. the goal of Prism is to start bridging that gap, and to achieve that goal they have provided a “browserless” browser interface in which you can run your webapps. it’s ultra-fast, clean, and each Prism instance provides its own cookies/cache/etc. the only vestige of a web browser that remains is the status bar at the bottom, but that’s just a good idea, as you’d like to see where you’re heading when you click on that link.

pretty much everything you need to know to get going with Prism is in their wiki (and there’s not much to know). give it a try and let me know what you think.

my to-do list of choice is Remember the Milk. unlike other web-2.0-ish online to-do lists, they don’t require you to visit their site to check your list (though you can if you’d like). they have built a Firefox plug-in that modifies the layout of your Gmail page to insert a collapsible to-do list right next to your email. you can add, edit, and complete tasks right from the interface — pretty much anything you need to do — and it’s all very seamless. if you’re in the habit of starring important emails, you can even automatically generate tasks for them.

anyhow, give it a try. it’s certainly helping me (until my brain finds a way around it, anyhow).

the next step was to find the ACPI control scripts for my T42 Thinkpad — or else write my own from scratch. luckily, i found the scripts required thanks to Aaron Tiensivu’s blog. just download this zip and follow the instructions in the readme file.

the default mode for the script is to turn on ACPI control through NHC, so that the fan is tied to the CPU and HDD temperatures (which makes sense, right?) — and the default temperature settings seem conservative. i have been monitoring the CPU temp closely to make sure everything is working as promised, and so far i haven’t been frying eggs on the keyboard.

the only odd thing is that NHC cycles through the fan speeds on startup and when you come back from sleep mode, so if you’re not prepared for that, it’s a little strange. the fan will rev up to top speed and then drop back to zero, which as far as i can tell is used to check that all the speed levels are valid.

in the end, the whole ACPI-script process took me about 3 minutes, and now my laptop’s fan is working normally and sanely.