Life is grand

The screenshot shows a performance improvement. From an F initially to a D after half-an-hour of work. ySlow is a useful little enhancement to Firebug which scans your web-pages looking for common performance issues such as too many JS and CSS files (external requests) and non-compressed files.

My above score is done off of my development machine which doesn’t have gzip, expire headers, etags and, obviously, a CDN. All of those except the CDN would be on the live version of the code so I expect a significant rating improvement for the live site.

The main tool I used to improve the score was AssetPackager, a plug-in for Rails which every Rails site should try.

Obviously we all need to be careful not to end up obsessing over ySlow scores at the expense of our users. ySlow is based mainly on Yahoo! sites and you may be building a site that is so different the same performance rules don’t apply.

Thanks to Chris for putting me onto ySlow.

Brendan Eich has announced three interesting script related projects at Mozilla. ActionMonkey you can read about here.

The two that are really interesting though are IronMonkey and ScreamingMonkey. The latter is a way to have the Mozilla JavaScript engine running in other browsers e.g. Internet Explorer. I assume this would be via a plug-in and it means we can have good JS performance and a consistent JS engine to code against in any browser ScreamingMonkey gets into. You simply attribute your script blocks appropriately and it will use the ScreamingMonkey engine.

But IronMonkey is the real corker. SilverLight had the good idea of allowing Ruby and Python to run alongside JavaScript in a browser. IronMonkey is that same idea but from Mozilla and “open”. I haven’t dug through all the details yet but I’m assuming it will, unlike SilverLight, be running Ruby and Python outside of the plugin and in the main DOM. So no bridging (correct me if I am wrong.)

Exciting times for web-development and these three initiatives will be very welcome.

Anyone else finding TechCrunch increasingly irrelevant? First the reams of “X acquires $ZZZ funding” posts and all those rumours and tumours. Money begets money and all that and obviously Arrington is enjoying the high-life which flows from VCs to him and back in a virtuous cycle of ever increasing hype, but from half-way around the world, and trying to get my job done, it is deathly dull.

Now TechCrunch is bringing us the blow by blow account of some daft startup that, would you believe it, blew through a bunch of money. That’s a first. That they apparently spent some of it on porn DVDs is obviously national news to some. For a bunch of guys who throw parties I am surprised they think porn DVDs are the worst thing companies spend money on. Of course shortly after posting TechCrunch had to strike out half the claims and add in a bunch of updates counter to the original post. Solid reporting here folks.

I’m putting TechCrunch into the “purgatory” folder of my feed reader. I’ll check it once a week as it limps along with its celebrity style blogging and if it doesn’t improve the unsubscribe button is easy to hit.

What a pity. When TechCrunch was just Arrington from his desk, without all the influence of money, parties, whispered tips and the demands of a money-making blog, it was good. The tone of TechCrunch was better then too, less sniping and cynicism, more love of technology and the exciting times the web is going through. Now TechCrunch reads like a celebrity pimping tabloid.

Apologies if I am a bit distracted today and don’t give you my full attention. I’ve got a final system test today then a tag rugby semi-final game to play this evening after which Fi and I have to drive to Dublin at 11pm to be ready for an all-day demo to O2 at 8am. Later that Friday we have to cruise down to Carlow for a wedding and a Saturday spent with Fi’s family. Hopefully I can then get back to Waterford on Sunday to be in time for sailing in Dunmore East.

Enjoy your weekend :)

A bunch of major sites have been hit by a power outage at their San Fran data center and it made me wonder what web-services I’d be OK without for a few hours. Thankfully all of the sites hit in this case are useful but not necessary. Had I been a LiveJournal, Vox or TypePad user I would be inconvenienced but not too put out. If I made a living off blogging though and it was hosted on TypePad I think I would be hurting.

At this late hour I can really only think of GMail and FeedBurner as being the sites that would hurt me if they went down. Though thankfully I’ve long used a redirect for FeedBurner so anyone subscribed to my various feeds can be switched back to the source and not routed through Feedburner. I’d loose some stats and a few nifty features for a few hours but I could live with that. With GMail I’d loose some emails but I’m assuming, maybe wrongly, that GMail doesn’t live in one datacenter and any system sending me email should retry a few times over expanding hours until the service comes back up.

If my site’s hosting went down then there wouldn’t be much I could do but it isn’t critical. You just wouldn’t be able to read my verbage for a few hours. Loosing Flickr for a few hours wouldn’t hurt and Twitter, heck, that could have been down for the past two days for all the times I’ve used it.

I would miss the Google Talk service but I’ve got a few other IMs and email logins spread around for rainy days. I read all my feeds through my own service, Millifeed, and thanks to our sysadmin and soon to be distributed hosting one datacenter collapse should only slow the service down a bit.

I’d probably really miss the daily Dilbert and Calvin & Hobbes cartoons though. That would be something you’d want distributed.

How about you? Do you use the web for a lot of services and how would you cope if a few of them went down for a few hours?

Martin Murphy linked to an interesting bit of spouting by a Nokia exec on how the Nokia N95 walks all over the iPhone. I won’t get into the feature race debate but will say that having both devices I know which device I’d rather carry around and it isn’t the N95.

What I will loose my lunch over though is:

For a mobile handset to be able to access Web 2.0 applications, that is the key to the future.

Barring the lack of 3G the iPhone is the king of “web 2.0″ delivering phones. The bigger and better screen, the better browser experience, the touch-screen interface. Navigating Flickr, GMail, MySpace, FaceBook, Digg etc. on a N95 is impressive until you try the iPhone. The N95 has a great browser on it (same internals as the iPhone browser actually; webkit) but you have to use the four-way scroller to move a tiny pointer to the link you want to click or to scroll down the page. Hold it down for 15 seconds and you might just get to the end of the Digg page. On some websites I have been unable to select a link because it is between the spaced hops the pointer makes (it moves up and down, left and right on a grid.)

On the iPhone you flick the screen and it scrolls. You tap the link you want to visit. You press the input field you want to fill. You double-tap to zoom in, double tap to zoom out. Want to pan North East? Just drag your finger.

Another thing for the Nokia exec to remember is the numerous iPhone tailored web-apps that are popping up. Digg has an iPhone URL. I haven’t seen one website offer something special for the Nokia N95.

So if the key of future phones is delivering the best web experience then Apple has only one step to make; add 3G. The Nokia N95 has to completely redesign it’s hardware interface and, having used Symbian applications, I’ll buy you a pint if you can make one have the same responsiveness as OS X on the iPhone.

Just a quick tip for Ruby on Rails; Don’t do render :text => 1 Rather do render :text => '1'

The former causes problems over HTTP when the client is expecting a content length of 4 but only getting 1. You get hanging connections which eventually reset. The latter render will send down the proper content length.

It seems obvious but thanks to the loose typing of Ruby you can, and I did, easily get bitten.

I’ve spoken about Structure101 from Headway Software before but they’ve just released a new version which adds in a very clever and useful feature. Where before Structure101 was a retroactive system it has now become a proactive tool through the ability to define your architecture and have warnings displayed during coding when your code breaks the architecture. This isn’t simple class checking or namespace collision detection but the ability to ensure your new lines of code follow the already broadly defined architecture.

If only there was a Ruby and JavaScript version I’d be buying a bunch of licenses. For now it is a Java-only tool.

Check out Chris’s post for more on it (I think his post does a better job explaining the feature than the press-release on the Headway site.)

I’ve got to say I really miss my daily dose of Kottke.

All the same, I hope he is lavishing Ollie with attention. Maybe a Kottke clone is needed…

Jing makes taking screencasts (screen recordings) dead simple. You select the window or area to record, do your recording, stop it and then either save it to your desktop or share it on Screencast.com. No projects, no encoding settings, no voiceover details or any other complicated setup. It is perfect for sharing quick recordings with co-workers or friends while Camtasia, made by the makers of Jing, is good for more advanced screencast production. Jing can live in your Mac menu bar or as a glowing orb near the top of your screen, ready to take a screenshot or recording at a moments notice.