Siu Lun

So so so… everybody would’ve heard that Google Wave is now officially dead, and Google would release the code-base as an open-source project.

No surprised to anybody surely. I’ll reiterate my views again, the problem with Google is that it is too technically focus. The downfall of Wave, was that the ‘real’/actual product/output is nothing new, nothing people don’t already do, yet it tried to make itself out like otherwise. Those who supported it, are nothing more than novice on the web. As harsh as it may seem, the fact that nobody bothered to use Wave is the signal required. Perhaps it would find some life after it is open-sourced at some start-up, but truth be told, the ‘technology’ that Google Wave has brought us can be easily and cheaply replicated / simulated.

Let’s break it down to the most basic form shall we?

Straight from Wiki:
Google Wave is an online software application formerly developed by Google. It was described as “a new web application for real-time communication and collaboration”.[1] It was first announced at the Google I/O conference on May 27, 2009.[2][3] It is a web-based service, computing platform, and communications protocol designed to merge e-mail, instant messaging, wikis, and social networking.

Ok, so right now, we use e-mails with our e-mail clients, IM with our favourite IM client, wiki in it’s own place, social networks where it’s supposed to be.

Google’s big idea is: wouldn’t it be nice to have everything on one screen?

The answer is quite a realistic, no.

We as human beings like to organise things, Google Wave, in a sentence to summarize, is like my room after a crazy night.

Essentially, the method of communication between the different mediums that exist today are very different. We write our e-mails like mail, not instant messaging. We use wikis mostly to look up stuff, not to communicate. We have social networks to stalk ppl who we used to know, seeing what they’re up to, appealing to the OK magazine type crowd, as well as self-publicity, we still use IM and e-mails to keep in contact with those real friends. Twitter is just a free form of SMS, with added functionality of a real-time RSS news feed.

As to collaboration, we use online collaboration tools specific to the application domain we’re trying to collaborate, not a ‘generic’ collaboration tool.

Do you see why these different mediums can’t be combined into a single entity now?

Not to mention the fact everything Google tried to do in Wave was to essentially trying to supersede everything that has been built and used so far, and they think they can just open the API to developers and it’ll self-develop. I tell you, developers would fare much better monetising each of the above communication medium on their own than to use Google Wave and essentially, giving Google a share of their pie they already had.

Finally, I would like to add that, Google anything other than search doesn’t sound good. From a branding perspective they ought to use something else. Google Wave, sounds like Microsoft Hello. Do you get my drift?

In this day & age, the most succesful services are usually not associated directly with a brand. You don’t see FlickR being renamed to Yahoo Photos now do you, and I bet when you do that. Yahoo Photos would just suffer a long slow death. Killed by the corporate brand.

Having said all these though, Google should’ve tried to think of ways to continue with Wave, rather than giving up.

I bet the story would’ve been different, had Google Wave been a core component of Google Chrome OS, much like Windows Live system on MS Windows, it would’ve garnered much more ‘public’ support.

Singapore AWS has been out the past couple of weeks, and I’ve migrated my server over. I have to say, accessing from Hong Kong is much faster.

If you also want to migrate your own server from US to Singapore or just want to use one of the pre-built AMI avaliable in the US but not yet in Singapore, follow what I’ve done.

First of all, start an instance of your AMI.

If the AMI you started is not your own, then customise it first, then execute within the instance:
ec2-bundle-vol -k private_key -u user_id -c cert

Once completed, upload the image direct to an S3 bucket in Singapore!
ec2-upload-bundle -b my-sg-bucket -m manifest -a access_key_id -s secret_key --location ap-southeast-1

Then all you have to do is register it. (commandline below)

If you have already got your own AMI, then SSH in a new instance of your own image, execute:
ec2-migrate-bundle -k private_key -c cert -a access_key_id -s secret_key --bucket my-us-bucket --destination-bucket my-sg-bucket --manifest manifest_path --location ap-southeast-1 --ec2cert ec2_cert_path --region ap-southeast-1

Your AMI will be uploaded to the Singapore S3 service, then you can use:
ec2-register --region ap-southeast-1 --name myvm my-sg-bucket/image.manifest.xml

Then you can launch your own SG instance! You may want to migrate any EBS volumes and data over. I didn’t have to deal with the hassle of that as there is no easy way to do it. I just backed up all files on my server onto my own machine and re-uploaded after.

I’ve been looking into building a sample Facebook Connect app recently for experimentation and showcase use.

One thing I encountered in recent projects is the seemingly slow speed of profile images loaded via Flash.

The one thing that I’ve found that is a potential ‘bottleneck’ to this problem is the way Facebook Connect is used on the site.

Essentially the fundamental difference from our recent projects is that we used PHP client side library which proxies all information into the server before being served to client side. Facebook provides a javascript based implementation as well.

Based on the above information you can basically tell that proxying is a speed bottleneck and an inefficient way to present what’s essentially the user’s information to the user only. Javascript solves the issue as it is a request from the client’s browser to Facebook directly.

Based on this theory, it should be obvious that php client library should only be used if the server has to manipulate, store or doing any information processing before the information is presented to the end user. Javascript library should be used whenever information should be presented to the end user first.

I’m in the middle of proving this theory. Will update once result is known.