Darren Steadman

Prevention is better than cure is one of those age old sayings mostly linked with medicine, well I think it also needs to be applied more in computing.

At current the computing industry thrives on cure, entire industries have been built on cure. In this case the cure Im referring to is applications that compensate for the inefficiencies or inadequacies in other software or areas of computing.

The problem I have with cure is much like in medicine the cure can often have undesired affects and only partially resolve the issue. Prevention is better because it stops the problem in its tracks and never lets it occur in the first place. The problem with prevention though is its hard to implement and often has no commercial or financial benefit when compared to cure.

A good example of cure is virus scanners. Here we have a world wide industry employing tens of thousands of people and making millions of pounds a year protecting us from inefficiencies in operating systems; They are providing the cure for a problem that would be much better served if it was prevented. Prevention in this case is a much more difficult problem to solve and even the current cures don’t always resolve the issue fully. Like cures virus scanners often have undesirable affects like slowing down computers or making a false positive for an important application or part of the operating system.

The prevention solution to replace virus scanners is a hard problem to crack. Many people argue that operating systems like Linux and OSX don’t suffer from viruses which isn’t entirely true and there is the argument that because of their relatively low market share compared to Windows that virus writers are less likely to target them in the first place.

I wrote an article a while back about the Mac app store and in that I put forward the idea that Apple could potentially lock down the OS in such a way that a virus would not be able to run at all. I then became aware some time after the initial announcement of the mac app store sandboxing which will be a further attempt to stop malware and viruses from running on the system. I see this as Apples attempt at creating a prevention rather than relying on the cure of virus scanners.

Behind all of this we also have a moral dilemma. If a company can modify their software to provide a prevention rather than relying on a third party cure should they do it and more importantly should they be allowed to do it? For example, say that tomorrow Microsoft announce that Windows 8 will be using some new revolutionary technique that means that viruses and malware could no longer exist on their systems should they be allowed to ship it? Now you may think I’ve gone mad here and you might be saying to yourself of course they should ship it who wants viruses and malware on their systems and who wants virus scanners slowing things down? Well removing the need for that virus scanner removes the need for that world wide industry employing tens of thousands of people and making millions of pounds every year; with the way the global economy currently is the last thing we really want to be doing is destroying industries over night. Don’t get me wrong I’m all for change and I think as technology progresses we will see a lot of changes in the world like this, but it doesn’t mean that the question shouldn’t be asked.

Another case of prevention vs cure can be seen from the recent articles going around the internet about turning comments off on blogs. I posted a reply to these articles in general, and it highlights to a degree what I’ve been talking about here, in this case using spam filters and moderation as the cure to a problem that needs a prevention (The article outlines a potential prevention).

There is of course the commercial argument for cure over prevention. There is more money to be made from cure than prevention. Making an OS that is already market leader no longer require a virus scanner is probably not likely to make that market share any greater, it might make the customer more happy with the product but it’s not likely to make any more money where as a virus scanner can have reoccurring revenue charging a customer a monthly fee to get the latest virus definitions to keep their systems safe. So this begs the question of is there any real motivation to do the hard thing of implementing prevention?

While I’ve only listed virus scanners as the main culprit here there are many other examples throughout computing, email spam filters, firewalls, tools to detect memory leaks in software, the list goes on and on.

I’m sure as time goes by more and more problems will move from cure to prevention as generally speaking prevention gives the user a better experience and creates a better more efficient system.

While the problem of prevention is a hard one I think it also creates more interesting and unique solutions and a certain level of elegance and efficiency that interests me greatly.

This is a reply to a post a friend made about a new technology he is playing with called FluidDB.

In his article he questions the privacy and data control issues with using such a technology as FluidDB, I found these issues interesting and so have decided to outline how I think they could be addressed. My thoughts do not deal with privacy and data access separately they kind of provide an overall approach that could potentially solve both issues.

From what I have read from both Guys article and some of the documentation on FluidDBs website, the premiss of their technology is that data is described as objects which are public to anyone else using the service, each user can then add tags to this object to extend the information about it or specialise it for a particular purpose. Tags can have various permissions allowing them to be readonly or restricting who can see them etc. All of this data is hosted in the cloud and anyone can access it from anywhere using simple RESTful type access using HTTP (I believe).

Now before I begin I would like to define what I class as “the cloud”. There are lots of different definitions of the cloud out there depending on who you talk to and what they are using it for. Personally I see it as distributed processing and data storage capability that is highly scalable, the most important part though is that you have a single point of interaction with the system and at no time do you know (or care for that matter) where the processing or data storage is. The whole point of the cloud as I see it is you have an application and you want to use that to interact with data of some kind of give a result, where that data is stored and where the processing takes place and more importantly how it is distributed is of no concern to the user, those are details that are taken care of by the cloud, all the user should be bothered about is that data goes in something happens somewhere and an answer comes out.

Using the above definition of “the cloud” I can then continue to define how I think the privacy and data control issues could be resolved.

Instead of having a single datastore, many service providers could provide data stores for you to place your personal FluidDB objects. It could be possible for you to even host your own FluidDB objects so that no company or other individual had access to it. Any FluidDB tags would then be held by the company/individual who was tagging your object and there would be tracker links between the two somewhat in the way bit-torrent works for finding peers. This would allow people to move their data around in anyway they wanted but still keep connections to tags etc. You could imagine that one day everyone would host their personal objects on a mobile permanently connect device such as a smartphone. Your personal object would contain any of the data you did not mind being made public and then various services could add tags to that object that could be marked as private and managed by the individual service you had signed up to. Obviously some of these services could provide public tags that other services could access.

In many ways it would work in a similar fashion to how google wave proposed the distribution of waves between enterprise customers to keep them private and secure while at the same time allowing access to certain outsiders.

This all means that data would be stored in “the cloud” (as defined above) and that anyone would be able to access it from anywhere but at the same time you would have more control over the privacy of your own data and no one entity would have direct control over all of the data.

This has been a very brief overview of a potential way to solve an issue that might not even exist, but sometimes it is fun to have ideas just to spark off some conversation.