Oracle EPM Smart Space

So for the past few posts I have been talking about blurring the lines between desktop and web applications and the technologies that will get us there. And in my first post on this topic I mentioned a technology called 'site specific browsers'. Site specific browsers are just that, it is a browser window dedicated to one site or in the case of BI and EPM a web based application. The installed application gets a shortcut on the user's desktop or the start menu and is a simple click away from running. You can read more about site specific browsers and how Mozilla Labs describes their product 'Prism'.

So in doing some searches I found that there are 2 major options with site specific browsers, Fluid and Prism. Fluid is Mac only so for my testing I used Prism. I decided to throw a couple of BI applications at it and see what I thought of the user experience. (I ignored minor bugs or issues as this stuff is still not official product). So I put Workspace (9.3.1) and OBI (10.1.3.3.2) in Prism and played around a bit. Here are some screen shots:

Workspace with a planning form loaded.

OBI with the default Paint Dashboard.

Both applications worked pretty well out of the box and it was nice not to have all the clutter and overhead that a browser carries with it. It was also nice to have the applications launch from a shortcut on the desktop. Some of the things I didn't like were as follows:

1. There were times where it would have been nice to have navigation (ability to go back). This is not an issue with Prism itself, it was the fact that the application was designed to live in a browser.
2. This does not change the application experience that much. (I did not expect it to) Basically if you are happy with the application experience this will improve on it. If you hate the way the application works then this will do little.
   

Overall this is a very cool concept but you are simply making web applications run as if they were desktop applications. I will use Prism for things like Google Mail or Yahoo Mail but at the end of the day, from a user experience perspective, I still prefer Thunderbird or Outlook for email. It would be cool to see if an RIA (Rich Internet Application) developed in Flex or Silverlight can run in a site specific browser…

OK I will be totally honest I don't have a whole lot on this one, simply because it is the newest entry in the market. It does look promising from the videos I have seen and what is being said about platform support sounds great. I just hope we can avoid all the issues I have seen over the years with JRE and version compatibility… I have signed up to preview the SDK and when I get a hold of it I will be sure to share more. I do want to share one cool feature I have seen on video. It is the ability to start with the application in the browser and then drag it to the desktop. Here the video that shows this:

The feature I am talking about is about 2:10 into the video and I think this will be a key differentiator that the other RIA (Rich Internet Application) players will quickly try to copy.

In my post on desktop and web convergence I said I would spend some time on each of the technologies I mentioned, so here we go. I will start out with Silverlight only because for the past 2 years I have been immersed in c# and smart clients so it was a pretty simple jump. Don't expect that after reading this post you will be an expert in Silverlight as I have only scratched the surface. I figure the best way to start is to list the pros and cons; here are some things I think fall into the pro category:

Language Support

Silverlight allows development and extensions using a verity of programming languages including JavaScript, C# and VB.Net. This will help ease the learning curve for existing developers familiar with these languages. Silverlight also has a Dynamic Language SDK that allows developers to communicate with the .NET libraries included with Silverlight. This makes it open to many other language possibilities like Python and Ruby.

IDE availability:

This is often time a con for newer technologies but for Silverlight this is a pro, it's integrated into Visual Studio 2008 it has a separate design environment of its own (Expression Studio). Studio is truly for the designer and allows for UI driven creation of animations and overall user experience. Having great IDE's for both the designer and the developer is a big plus.

Platform and browser support:

Silverlight supports many platforms and browsers. This is different than one might expect from Microsoft but Silverlight supports IE, Safari, and Mozilla and will run on both Window and the Mac. Plus there is now support for a growing list of mobile devices.

User Experience Support:

Silverlight has a wide variety of features that support a rich user experience, this combined with the power of Expression Studio make for crisp UI's. Features include; media support, panel and canvas support, animations with timelines, AJAX support, etc.

Here are some things that I considered cons:

Limited .Net Framework:

This one can go either way but for me it is a con. Silverlight comes with a limited set of the .Net framework assemblies (to keep the client install small) and for developers that are used to desktop development this will be difficult. For others new to the .Net world they this will not be a con.

Separate IDE's for Design and Development:

I am always looking for that one IDE that does it all and I tend to cross the line between designer and developer so switching in and out id Visual Studio and Expression Blend was not so smooth. I wish they would just stuff Blend into Visual Studio but I am sure many disagree on this.

Keep in mind that Silverlight 2.0 is currently in beta and like any other beta software you should use and install with care. Here are some cool Silverlight samples.

Just as I was thinking about what technology I should start with from my last post another company decided it was time to hop on this desktop/browser train. Sun announced JavaFX at the JavaOne conference yesterday. JavaFX is the latest entry into this space and this should help to fuel innovation but will add to the confusion when trying to decide what platform to develop on. This technology along with the others has many compelling features like rich user experience, browser and desktop capabilities, and it's cross platform.

Stay tuned next week when I start to dig into each of these technologies...

This is not a new topic but it is certainly new for BI and EPM. For the past 5 or so years the focus has been on providing users with functionality while attempting to have zero or limited client deployment. In other words build your front end to run in a browser, preferably with no plugin's (HTML and JavaScript only). Unfortunately this has become more difficult as users have demanded more interactivity and performance from clients. They get this from the consumer application space; while in ITunes they use cover flow to select a song or while checking stocks with a vista gadget the chart animates open and closed. Really we could just select a song in a list view and see the stock chart in a popup window but that would not be cool nor would it differentiate the application from a competitors'. Let's face it innovative user experience is hot! Lucky for us there is now some great technology to help us meet the needs of our end users while still allowing for simple application deployment. Here are a few:

Google Gears

Adobe Air

Microsoft Silverlight

XBAP –XAML Browser Application

Microsoft ClickOnce

Site specific browsers like Fluid and Prism

To try to cover these in one blog entry would not do any of them justice, so stay tuned…

I came across an article in Hemispheres Magazine (UAL publication) about this time last year. The article was about cycling, and last February I was just getting back into cycling after a long break from it, so I was intrigued. Aside from getting back into cycling I was also buried in everything Smart Space attempting to build and launch this new product. I was shocked how much this article related what I was doing in software to what Shimano was planning to do for cycling. The article started off talking about how the Lance Armstrong era of cycling was over and the wave of high end bike sales was coming to an end in the U.S. So Shimano went off to figure out how they could get people to again love the sport of cycling and therefore buy more products. The conclusions they found is where all the similarities came. They found that people had very fond memories of cycling and really enjoyed the idea of it but could not wrap their head around the complexity of what cycling has become. Things like 10 speed cogs, carbon forks, STI shifters, and clip less pedals had started to overwhelm the average cyclist. The basic conclusion was that people just wanted to ride a bike; simplicity. This is what I really related to; the goals of Smart Space are in line with this, end users want simplicity. Many end users don't want to worry about connection settings, indexes, caching, dimensionality, etc. they just what to know how their travel expenses will impact the department's budget for this quarter or get the status of their sales pipeline. I saw that the key for Shimano and Smart Space was in simplicity. So whenever I think about a feature or plan out a new gadget I think about the idea that people just want to ride a bike.

Here is the article so you can decide for yourself if the concepts really relate:

http://www.ideo.com/pdf/Hemispheres_2007-02.pdf

When talking to people about Smart Space I hear this question come up all the time. I have found that most people have very different views on this topic so take what I have to offer as merely another opinion. In my earlier post I talked about the definition of a gadget and stated the following:

Gadgets (or Widgets) are mini applications that expose key content (bits of data) or features generally from a larger (full) application and they deliver these features or data in a simple and visually pleasing manner.

So when I read this I key in on some key concepts that help me to differentiate between an application and a gadget. First, I wrote that gadgets are "mini applications", and to me this means that they are smaller than an application and, at times, related to a full application. When I say smaller I mean smaller in two ways, smaller in physical footprint and smaller in the screen real estate that the gadget takes up. Second, gadgets focus on "key content"… "or features" where an application will have many features and tons of content. Lastly the gadget should present this information in "a simple and visually pleasing manner". In other words when a gadget is giving me information I should not have to guess at what it is telling me, the presentation of the data is just as important as the data itself.

Here are a few examples:

In Smart Space there is a nice search gadget that lets me search for content in Hyperion Reporting and Analysis (System 9). It is very simple, just enter a search term and get results. This is the kind of search I do 99% of the time and that is why this makes a great gadget. If I want to get more advanced I could open the Hyperion Reporting and Analysis application in my browser and navigate the search page to perform the search with a number of other key options. The gadget takes up very little room on my desktop and covers most if not all of my Hyperion Reporting and Analysis search needs, but the application is there when I need it.

In our beta I wrote a notepad gadget that is great for taking quick notes and having them always visible on my desktop but I would not want to write this blog entry using it. For writing emails, document or blog posts I want to use an application like Word that is full of great features for writing.

In the Smart Space Key Contacts gadget I can limit my list of users that I communicate with, down, from the long list that includes people I have seldom contact with, to a much more focused list. At a glance I can see who is available to chat, and with a single click I can start my IM application. In this case the gadget provides visual indication of my key contacts that are available and launches me from the gadget experience to the application experience.

Here is an example from the consumer gadget world and this should drive my point about presentation of the data. I will use images to demonstrate this:

Both deal with system monitoring but the gadget gives me the basics and at a glance tells me what I need to know. (My CPU is fine but memory consumption is a bit high) If I want features and details then I go ahead and open the application.

To conclude I want to keep things simple, so when creating a gadget don't try to satisfy every use case otherwise you will have an application on your hands, make sure that you are building something a user wants to run on their desktop all the time, and make sure what you present has the right design for a user to 'get it' at a glance. I have found that these same Ideas can be applied to almost any application and I think about these concepts whenever I am building a new gadget.