API Evolution (Page 62)

Writings about API Design, OpenAPI, HTTP and related topics.

Google versus Microsoft. Phooey!

I'm guessing it is a slow news month. Everyone seems to be talking about how Microsoft better shake themselves up or Google will beat them up. Forgive my candor but what a crock. Let's look at what Google do well. Their search engine is way better in my opinion than Microsoft's. I don't know what MS are doing but their own internal MSDN search engine is one of the most useless I have ever come across. How they expect to index everyone elses content if they cannot sort out their own I don't know. maps.google.com is phenomenal. It kicks MapPoint's butt. Gmail kicks hotmail's butt! There is no doubt that Google has developed some really cool stuff in the web app space. I like Google's UI philosophy. Make it simple but think about it a lot. You can tell that somebody spent a huge amount of time making Gmail and maps as effective as they are and yet the end result looks very simple.Interestingly if you check out www.start.com you can see that Microsoft are starting to get the idea. Much of the premise that Google have the potential to overthrow Microsoft is this idea of the Web as a platform. But hold on a second, Google's last few offerings, Google Earth, Google Desktop search and Google Talk are all client applications! Sure they are internet aware apps but fundamentally they require an OS to install on. As a sidenote, it would be interesting to know how much revenue Google are making on these tools. I have read some comments on newsgroups that Google are planning a Linux based OS. Of all the Unix distributors, I would say that Google has the best chance of pushing a Linux variant into the mainstream. This is certainly a much more realistic option than the pie in the sky web platform hoopla. If Google do go that way they are then going head to head with Microsoft trying to deploy a client OS. I don't think Google will survive that fight. Microsoft have the money, the people, the marketshare and they can actually charge for their OS. Google are an innovative company, with some great web apps, that has a very clever way of dealing with the media and its own image. They don't have an OS, they don't have an Office system, they don't have a huge network of partners, they don't have business applications, entertainments systems, they don't have developer tools... I could go on.

Read more...

Prediction time…

I have always enjoyed making predictions about where the IT business is going. Now I can blog about it and realize in a few years how wrong I was! IBM is going to buy Sun. The main reason is to prevent Oracle from buying them. Oracle have the database, the application suite, they are talking about buying BEA which would give them the application server. Buying Sun would give them hardware, an OS and control over a programming language that they are already dependant on. Buying Sun would give Oracle a complete end to end platform, just like Microsoft have. IBM have bet the shop on Java and I don't think they can risk Oracle getting control over it.

Read more...

MBS is headed the wrong way…

Call me a skeptic, but I just don't think MBS are headed the right way with their new roles based model for Microsoft Dynamics. In a number of the recent announcements made at the Business Summit and PDC, they have talked extensively about the 50 roles that they have defined to try and identify the work performed by the medium-sized business workers. There are two major problems that I see with this. First, the majority of workers in the majority of companies will not quite fit in any one role and second, the tasks defined for the roles will only work for a subset of the scenarios that the workers run into. In this global marketplace, unless you are a physical storefront retailer, there are very few identical businesses. Companies need to have a differentiating factor or they will not survive. This uniqueness of businesses can put a twist on almost every aspect of the business. Sure, depositing a cheque in a bank is conceptually the same for a grocery store as it is for an engineering consulting company, but I am quite sure the internal process for handling those cheques is very different. Creating an invoice at a bolt distributor is not the same as creating an invoice for a pipeline manufacturer. I guess what I am saying is the vertical focus is absolutely critical to the process flow. Not only that, many companies are not homogeneous. The invoice printed today for product A likely has very different constraints than the invoice I may print tomorrow for product B or even service C. This is the point where MBS would say, "that's where our partners come in. They have the vertical knowledge and they can configure the processes to match the needs of the specific business." Maybe. How many companies have the time, money, expertise and desire to extract the details of their process flow for all scenarios from all of their knowledge workers? Sure it is an ideal scenario. The ERP system knows exactly what is the next step to take in every scenario in every business process that can occur. Well at least until it changes! What I see happening here is analogous to the debate between using MRP II to manage production and the lean manufacturing principles. The MRP II process attempts to use a computer to manage every last detail of the material requirements and procurement process and it can become very cumbersome and fragile. It works in an extremely controlled and homogeneous environment, but the reality of many companies is far from that. I believe we should be trying to develop tools that empower knowledge workers to do their process in whatever way is the most efficient for the specific scenario they are dealing with now. They understand their business, they understand what needs to be done, they do not need to be lead by the nose through a predefined sequence of steps that was implemented by an external consultant. We don't need a process defined on top of Outlook that says 1 - Read emails from Boss, and reply if necessary2 - Read emails from colleagues, and reply if necessary3 - Read joke of the day, if available4 - Repeat every 30 minutes We each have our own internal process that we have developed for managing emails. Sure some people are better at it than others, some people get way more email than others. God help the consultant who tries to tell everyone they have to read their email in exactly the same way at all times. My experience with knowledge workers is that the standard scenarios are not the ones where productivity is lost. It is the exception cases that cause the problems. We needs to business applications that sit there and say "What can I do for you?" "What do you need to know?", not "Let me show you how to do the work you need to do today." I am not saying the Role/Task based approach does not have its place. If I am an occasional traveler and I am going on a business trip, sure I want to be lead through the process. If I want to sign up for the company pension plan, I want to have my hand held. If I am doing my day to day job, no computer is ever going to tell me the best way of doing it, even if a legion of consultants analysed my company for years.

Read more...