The bottom line is that a well run cloud will be much more fault tolerant, robust and scalable than anything an individual or a company could install while at the same time being more economical.

Personalisation has long been an important aspect of knowledge workers – to arrange knowledge and information in the way that makes sense to the user so he can work in the most efficient way.

Good knowledge managers have also known that enforcing a way of working that matches the corporate way, the taxonomical way, is also bad because most of the time “my way of working” and “my mental model of information” doesn’t equal “the corporate way”.

Email is a good example of a great app that allows for high levels of personalisation. Fortunately, the corporate taxonomists don’t enforce a structure that users need to keep their emails in.

Enforcing a cloud view, however, is probably as bad as enforcing a taxonomic view of the world. It’s still only one way to look at a user’s own information, but it probably won’t be the user’s preferred view. The biggest tag may not necesarily be the one that they need to keep tabs on today or for the next month. What if the user has thousands of tags?

IMHO, discussing the need for folksonomic and taxonomic views are important, but the ability for the individual to organise information in the way that he thinks, to be able to personalise information so that it reflects what ever way they work, is the important factor for all knowledge workers.

I posted on this a few weeks ago, asking if it was time for an “Internet File System” to complement the emerging Internet OS’s and allow for separation between applications and storage:

http://techfold.com/2007/04/18/time-for-an-internet-file-system-ifs/

Check out that mocked up Google Docs “Save to: .mac, box.net, etc” screenshot.

http://techfold.com/2007/04/18/time-for-an-internet-file-system-ifs/

Check out that mocked up Google Docs “Save to: .mac, box.net, etc” screenshot.

….I thought about what it would mean to work 100% in the cloud. The scenario that I bumped my head against was that unless you did all of your work inside a single “OS” (Google Docs & Sheets, for example, or on apps within the Xcerion platform), managing files and work scattered across the net, behind different passwords, formats, export capabilities, privacy policies, and more, would be a severe pain.

That is to say, what happens when you use Google Sheets, Zoho Writer, and Picnik on a daily basis to get things done, and Box.net as an online repository?

Perhaps I’m old school, but whether my data is in the cloud or in local storage, I still see value to centralization:

1. Being able to search and index all of my files
2. Being able to organize my files topically – i.e.: a spreadsheet for trip expenses, photos, letters, etc. from a trip all stored together (whether tagged, or in folders) as opposed to scattered across multiple services.
3. Not depending on the solvency of multiple companies to take care of my data
4. Being able to run a centralized backup function on all of my data (ironically, this probably means backing up my cloud storage to a local harddrive)
5. Being able to open a file created in one application in another (i.e.: opening a Word file in OpenOffice == opening a Google/Writely file in Zoho Writer) without a cumbersome import/export process

So – what I think I’m suggesting is an “Internet File System” that would let online applications save to and load data from a third party storage solution. This would include:

1. A Protocol: A standardized protocol for applications to implement on their end, and storage providers to implement on theirs that would manage the connection seamlessly between the two. The protocol might include two-way communication so that I could see how much space was left in my Box.net account from within Google Docs & Sheets.
2. Universal and Free: Note that this is a “protocol” not a product – it should be something like SOAP that anyone can implement. And FOSS’d.
3. Encryption and Authentication: I’d like to be able to give Google my Box.net password once, and have Google manage authentication with Box.net from that point forward. I’d also like the datastream between application and storage to be encrypted.
4. File Meta-Data Standard: Some sort of meta-data standard for the file objects that would be saved such that when looking in Box.net, I can click a spreadsheet file and have it open in Google Sheets where it was created.
5. Format Meta-Data Stardard: A central repository of DTD’s/XML Schema descriptions for data/file formats, to allow for easy conversion of file types between applications. I don’t want to be locked into a particular application by its format. Format owners retain ownership of their format (and could GPL it if desired), but regardless of ownership it should be available in parseable format for other to build converters around.

Ok – that’s a tall order to be sure. I have trouble imagining many companies wanting to give away control to that degree, and to that point, this post is more to plant a seed or spur discussion than anything. Plus, as noted, I may be in conceptual never-never land and all of this sounds really dumb.

At the end of the day, though, I want to jump into the cloud. I’m ready. I just don’t think the cloud is yet.

I’d like to see this become a checkbox in WWD’s reviews of web apps – do they back up customer data. For a free app (Gmail, Yahoo Mail, etc) this is a nice to have, but for a paid app this is a must.

Also, there’s a difference between an outage and data loss… If Gmail is out for a couple of hours it’s annoying but if they lose my data it’s catastrophic. Which is why I do have all of my mail sent to my domain and forwarded to GMail. It then exists in 2 data centers and the odds that I’ll lose both are so low that I can’t worry about it.

This reminds me of a hard lesson learned during a computer programming class in junior high school (some many, many years ago). We all sat at computers lined up on a long desk with all power connections controlled by one strip at the end. One day, the instructor walked over and turned off the power strip (without any warning) and we lost everything if we had not saved our work recently. Back then, it took hundreds of lines of code just to draw a square, so I was mortified.

Now, my fingers always automatically hit CTRL S to save what I am doing and all of our servers (in-house and customer databases) are backed up religiously. Thankfully, we don’t have much of a need to retrieve data because our up-time percentages are very high… but if that day ever comes, we got it covered.