by Jon Poulton

One of my favourite stories about London is about the Millennium Dome and a visiting French President. After being shown around the dome just before the year 2000 celebrations, the President looks squarely at his guide, and says "Yes, very impressive... but what is it for?". This was almost my exact reaction upon seeing Google Wave for the first time. It looked pretty, and the Google employees presenting the technology seemed reasonably enthusiastic, and it seemed to have some neat features. But I was still left wondering what on earth it would be used for. Is it a chat tool? A replacement for email? A new kind of discussion forum? A tool for collaboration? I've read a number of articles that have touted Wave as all of these things and more, but I'm not entirely convinced by any of them. Let's try looking at them one at a time.

A chat tool?

Right, so it's a chat tool, with a browser interface and lots of nifty gadgets that allow you to do things like insert maps, images and add software agents to the chat. Cool! Well, cool except for the fact that it has a number of features that I don't really want to see in a chat tool. As a chat tool, it doesn't have features, it has anti-features. What's sad about some of these anti-features is that, from the perspective of a programmer, they look very interesting and obviously took a lot of effort to implement.

Our first anti-feature is the live-update of text within Waves. I do not ever want anyone to see what I am typing before I have finished typing it, and decide that it is ready to be seen. This includes things as small as chat messages and tweets. Text on a web page or within an email has no vocal tone. Something you would normally be comfortable saying out loud, in company, can end up looking sarcastic, arrogant, dumb, or devoid of the humour you intended it to have. For this reason many of us quickly and silently read back our own text before submitting it for others to read, subtly rephrasing what we say to best communicate what we mean. In part, Wave denies you this, as anyone watching the Wave will be able to see your clumsy attempts at wit long before you've actually finished typing. In essence, I just don't see the benefit of this feature.

Our second anti-feature is that anybody in a given Wave can edit any of the text that I have posted in that Wave. Someone obviously thought that having a collaborative wiki-like editing feature would be a good idea, but it makes no sense in a chat context. I don't want people to have the ability to edit my posts after I've already posted something. It's simply confusing to log on and see your post containing things that you never put in there. Wave does have the ability to "play back", edit-by-edit, every change that has been made to a given Wave. This does allow you to track who made what change and when, but it's rather long process for Waves with many posts and edits.

Lastly, our third anti-feature is a little more subtle. Say I'm having a conversation with Alice about Bill. Bill is having his annual review, and we've collected feedback from all of his peers and put it all up on a Wave to go through it and pick out how best to give Bill the feedback. As we're chatting about this on the wave, either myself or Alice accidently invite Bill to participate in the Wave. Now what do we do? Bill can now see everything everyone has said about him. We can't kick Bill off the Wave, because the interface won't allow you to do that. We can't delete the whole Wave, because you can't do that either. We could edit, or remove, the confidential data before he had a chance to read it all, but at any time Bill can always use the "playback" function to see what was there before we edited it.

A replacement for email?

One of the first pieces of hype I heard for Wave was that it was "what email should have been". I do not understand this assertion. Modern email has sophisticated methods for classifying, tagging, and searching for mail; for managing multiple identities; and for filtering out unwanted content. It's a simple system for sending text documents asynchronously between two or more people. In contrast, Wave has a folder-like system for classifying Waves, and a simple single-box-input search function. Wave does not make it easy to show all the messages between two users, or all the messages sent after a given date, because things are not organised in terms of individual messages. Waves which are "open" do not appear to have a way of removing unwanted content, or users, permanently.

A discussion forum?

Wave does have a few forum-like features, such as the most active waves being boosted to the top of a "most active threads" listing. Except it's not a thread listing, it's your "inbox". It does allow you to categorise in a folder-like structure, which I suppose is a little bit like using sub-forums. However, as you are limited to viewing and searching only those discussions you are invited to, or know of and are open, I think it would make a poor replacement for forum software.

A collaboration tool?

I think this is more like it; at VYRE, we are currently using a series of Waves to throw around new technology ideas and post links to each other. So far, it has provided us with an interesting way of collaborating in an unstructured way, and I think this is where Waves' real strength lies. However, I am aware that the sentence "Wave is a new form of collaboration tool" sounds far less impressive than some of the marketing hype that's been sent our way over the past few months.

I'm really looking forward to seeing what sort of agents are written for Wave, and how people end up using it in the long run. My bet is that it will be used by some people as a persistent form of chat, and for occasional pieces of informal collaboration. It is not an email killer, it won't replace chat clients, and I'm afraid forum software will be around for a long while yet.

On a final note; I would have preferred Google to have produced a specification and reference implementation for Wave, and allowed Software Engineers such as myself to play around with writing software that implemented the 'spec, giving us the opportunity to add our own touches. But this, alas, is not the Google way of doing things; and with the recent advent of "Google Buzz", it looks like we're still on track to see a myriad of browser-based Google-provided services, complete with revenue-raising advertising based on the content that you provide them. Surely there is a better way of making an honest buck?