Blogs

We got it done only 2 days late and minutes before I had to hop on a plane to go over to London one last time (if anyone wants to meet up and chat AJAX / Web2 then drop me a line).

As for our new AJAX Grid … if you are a customer or not I would love to hear people’s feedback on some of the ideas that I mention here. The EBA Grid V3.0 Alpha (video here) is meant to provide our current customers with a better idea of where we are going with our new and improved AJAX product line. If you have somehow heard about the Alpha don’t think that you couldn’t live another day without seeing it then email us and we will see what we can do.

There are three primary areas where we are extending the EBA Grid functionality. First and foremost we are providing support for Mozilla based browsers - the Alpha version of Grid V3.0 has been tested (I use that term as loosely as possible for an Alpha) on Firefox 1.5 and Internet Explorer 6.0 (Windows only at the moment). We are planning on officially supporting more versions of Firefox and Mozilla with the final release and will be looking for customer feedback to determine if we want to support other browsers (in the near future) such as Safari and Opera. From our point of view, the reason for not supporting most other browsers is because a) we use both XML and XSLT extensively in our products and b) our customers and the market are not demanding other browsers to be supported.

Secondly, the Alpha version will showcase our new Live Scrolling paging mode - Live Scrolling allows for users to use the scroll bar as though the data is there but the data is actually retrieved and rendered (using AJAX of course) only when the user stops scrolling. In the final version we will support a few different types of paging such as Traditional, and Open Ended (Live Scrolling or Traditional on arbitrary number of records).

Finally, the Alpha also includes a very green version of a .NET 1.1 / 2.0 version of our component. There is nothing too fancy and no designer support at the moment but we should have a nice .NET backend for the final release.

Our minimum requirements for the final release of our AJAX Grid V3.0 is to have all the functionality of V2.8 as well as support for Mozilla/Netscape/Firefox, Live Scrolling and a .NET server backend. There might even be an Alpha version of the JSF backend by that time too!

Other notable features that will be coming in the Beta and final release of Grid V3 include:

Backwords compatibility
EBA Grid V3 Alpha is a first attempt at building V3 backwards compatible with V2.x. Is this important to you for the future? If so be sure to let us know and we will work even harder at it - otherwise we will work on other cool features instead.

Copy and Paste
One of the other cool features that we know a lot of people use (and hate to use) is copy and paste. We have not included it now but our new copy and paste functionality is super fast so no more waiting for your data to paste!

Localization
The new Grid will support multibyte character sets for foreign languages although this has not been tested in the Alpha.

Debugging
Like a lot of the new features, we are not quite ready to release the Debugging interface for Grid so when evaluating the Alpha we suggest using MS Fiddler (IE+FF) or FireBug (FF).

Architecture
That being said, there is going to be a lot of advancement that you, the customer, will not necessarily see on the front end. The architecture for Grid V3.0 is going to be the basis for the rest of our product line as we move forward this year.

Component Oriented
One of the driving forces behind our new architecture is to be able to build truly component oriented applications in the web browser and to enable developers to easily extend the functionality through either JavaScript? prototyping or through XSLT. More on this soon!

Declarative Programming
Likely the most interesting part is our move over to a more rich Declarative Model for our components. Our declarative markup is one area where we are very excited to get some customer feedback as it is surely going to not only make our components better but will also be driving internet development over the next few years with technologies such as XAML, XUL / XBL, XForms and Flex really starting to catch on.

DataSource
Another interesting feature that is behind the scenes of our AJAX Grid V3 are a new EBA DataSource component which encapsulates all the XML data manipulation such as saving data to the server and merging updates from the server. We want to release this shortly on its own as a simple client side data management solution (licensing TBA). Furthermore, the data management layer will have other important AJAX functionality such as client side data caching as well as predictive data fetching.

Grid V3 Alpha Specifics

Known Issues

• Internet Explorer memory leak (don’t worry this will definitely be resolved )
• Some XSLT files are not cached by Internet Explorer.
  - DocType DTD. What happens in Strict Mode?
• Keyboard navigation can get confused with frozen rows and columns
• Various problems with live scrolling (column resizing, insert / delete record etc)
• Lots more

Other New Feature Ideas

• Record Key generation - how should keys for records be generated?
• Declarative markup - what is easiest for people to read?

New Feature Requests???

For those who have not read this already, there is an interesting article from Tim over on the Radar about NetBeans vs Eclipse popularity.

We are currently working on our Java strategy for our AJAX components and so this is a very big question for us. Having been discussing some of the possibilities of AJAX and JSF with the Java Studio Creator team at Sun. Of course JSC is built on NetBeans 4.1 and from the sounds of the comments in Tim’s post Sun may have made a good choice. From an AJAX component vendors perspective I think that JSC makes a lot of sense compared to Eclipse since, as Robert Thornton mentioned in the comments of Tim’s post

Eclipse only recently left the beta stage in its support for web development, almost as if it were an afterthought. NetBeans has supported web development for years as part of its core functionality. Its web support has had time to mature and grow.

In particular there are some really great features in JSC such as the DataProvider architecture and JSF support. I still need to find some time to actually play around with JSC some more.

Today a few of us at EBA had a sneak peak at some really cool Flex technology from the team over at Adobe.

Essentially we saw how one can go beyond the ExternalInterface feature in Flash 8, which only supports function calls across the JavaScript / ActionScript boundary with primitive objects, to a situation where one can access Flex objects directly from JavaScript and vice versa. This is being called the Flex-AJAX Bridge or FABridge. All the info (with Alpha download) can be found here on Adobe Labs.

Today this means that AJAX developers can better leverage native Flash capabilities such as local storage (hopefully it is faster than ExternalInterface), cross-domain data access, and sockets. It should get really interesting when the AJAX Client for Flex Data Services become available (later this year) which should provide data persistence, pub/sub, push, etc.

As a component vendor, I like the idea of marrying the benefits of both AJAX and Flex - it particularly makes sense for doing anything with charting etc. With the SVG/VML/Canvas battle for vectors graphics dominance pretty much going nowhere (and they can’t even do video or audio), Flash is the best option for rich media that works on a large majority of today’s web browsers. It was a good move on Adobe’s part to provide some incremental benefit for AJAX developers rather than trying to push a full Flex framework where the open standards of the AJAX technology stack rule. Something like XAML, which is Microsoft’s next generation declarative user-interface language, will have a tough time for that exact reason.

Flex seems to be shaping up into something that should, in the near term at the very least, be a great addition to the AJAX developers toolbox. I can’t wait to get into the guts of it - just as soon as Grid V3 is done

Andre, Alexei and I had a chat with Coté and Stephen from RedMonk yesterday. They both seem (and no doubt are) very knowledgeable about the AJAX space - certainly compared to other analysts with whom we have spoken (and shall remain nameless). Not only did they seem genuinely interested and understand the business problem that we are trying to solve but they (thankfully) didn’t even try to sell us their services! They seem like the kind of people we could work worth as we grow our business in the coming year.

What they understood was that the real value of AJAX is quite boring. It is in the mundane spaces like ERP - not the exciting, and ultimately fairly useless, arena of something like online maps. I have been talking about this idea of boring AJAX for some time and maybe people are starting to see that those enterprise systems like CRM are where the real benefits of AJAX will be found. Needless to say, that is precisely the market that we are targeting with our high performance components.

Another little tidbit for everyone - don’t use element.focus() when you have a lot of HTML elements on your page.

This is quite closely related to my previous post about using a:hover.

We thought we were sooooo smart - well actually James-no blog-Douma thought so (of course he has some right to think that since he is uber smart).

Anyhow, in the latest version of our Grid we thought that we would use <A> tags for each cell in the grid such that we could take advantage of using things like A:hover pseudo classes in CSS and the ability to set the focus on the cells - all this we assumed would be fast since it is built into the browser and does not require so much manual labour with setting active colours etc. Low and behold we found that performance of the keyboard navigation started to drop fairly rapidly as we rendered more cells on the page.