Smarter Everyone, Smarter Everything, Smarter Everywhere
John M. Boyer 060000VMNY 809 Visits
Very soon, I'll be heading off to the W3C Technical Plenary.
Sure, it's in Cannes, but it's also in February, so it's all business.
This is a time when many W3C working groups congregate in order to have time to have intergroup meeting times. So, the XForms working group will have its face-to-face meeting later in the week, but so will a dozen or more other groups. Similarly, the compound document format group chaired by Kevin Kelly will meet earlier in the week. I plan to sit in on some of that as an observer and to talk as needed about the intersection of XForms and compound document formats.
The mid-day of the plenary is a time for common presentations, one of which is a panel presentation called "Rich Web Application Backplane". Kevin is the chair of the panel, which has a number of panelists including yours truly.
The panel is basically about the notion that we have lots of rich XML processing capabilities available now, expressed as XML. This allows for compound documents that are capable of aggregating the data, views, rich user experience, and web services needed to complete a sophisticated web application transaction.
While some further work can be done to do a better job of creating more seamless integration of various components in the W3C stack, a lot of the foundational technologies are already there and in pretty good shape, so herewith is the rallying call to pull them together into an XML backplane capable of supporting the design of rich web applications.
One of the things I really like is the potential for XML events to be the glue that binds together nicely modularized XML technologies into an effective, intelligent compound document. I think an update to XML events would be useful to add certain key features, but even as-is the capture and bubble phases combined with event context have allowed me to express an integration of XML signatures into a compound document format that includes XForms and a first class presentation language (XFDL, of course).
Of course, whenever we talk about advancing the conceptual model of the web in some way, we get a certain amount of feedback about whether we are talking about evolutionary change or revolutionary change. The latter causes IT folks ot go weak in the knees over deployment issues, so branding something revolutionary is really about trying to stop it from happening.
This means that you can have a staged migration to a rich native client that directly runs the high level language because the conceptual model expressed by the language can be delivered, with few limitations, to today's client platforms-- zero footprint.
In the IBM Workplace Forms space, we call the native client "Viewer" and we call the zero-footprint version "WebForm Server". The idea is that you design an XFDL form containing precision presentation, rich user experience and business logic/dynamic interactive capabilities using XForms within the XFDL. Then, you deliver it using one of these two interactive rendition engines.
A demo of this capability is how my panel segment will end. After all, nothing trumps being able to show some of the kinds of capabilities that would be enabled by the rich web application architecture.
John M. Boyer 060000VMNY 808 Visits
Well, I haven't blogged for the last couple of weeks because, as it turns out, I was swimming with the sharks!
Well, actually, I tried my best not to be in the water when a pair of hammerheads and a trio of grays came callin', but one morning my wife and I had to rush into the water to help an older man who was merrily floating along, ears below the water line and a pair of hammerhead fins vectoring his way.
Otherwise, a peaceful and restful vacation and back in the XForms saddle. And while I was away, the W3C publication process did indeed unfold as expected (with a little help from Steven Pemberton) to allow delivery of the latest working draft of XForms 1.1
Notice that the above link takes you to the diff-marked version of the spec. These show the differences between the current working draft and its predecessor, which was published in December of 2005. If you have familiarity with the prior spec, this allows you to see the new things that have made it into XForms 1.1, and the changes to existing features of XForms 1.1
There are quite a few smaller tweaks here and there, but I'll focus a bit here on the big ticket items. One is the effort the XForms team has been putting into the access that XForms authors will have to event context information in XForms actions. This is going to be a very powerful addition to the language. The other major area of change involves the addition of features to the XForms submission module. The goal of these changes is to open up XForms to a wider variety of sympathetic web technologies. Frankly, there is an interplay between the changes we made to submission and to event context. Expect some further changes and additions in the next working draft.
A small change that has, perhaps, the most important and far-reaching implications is the namespace URI that will be used for XForms 1.1. This is a very challenging issue, but the net is that XForms 1.1 will use the same namespace URI as XForms 1.0. Like those sharks, this really bites in some ways, but also like the sharks, it turns out to be unavoidable. Stay tuned, and I'll tell you why in an upcoming blog entry.[Read More]
John M. Boyer 060000VMNY 806 Visits
An interested reader asked "How do you scroll a repeat?" so that it's content doesn't visually spill onto other controls. The question was specific to the Workplace Forms product, but the answer is reasonably generic.
One answer is that when you put things below a table (the XFDL element that contains and presents an XForms repeat), don't use absolute positioning. Use relative positioning. Use an itemlocation that puts the content "below" the table so that as the table grows, the content below it automatically shifts down.
The above suggestion is useful if you are allowed to the let the content grow arbitrarily. However, if you are trying to match the appearance of a paper form, then the table/repeat must fit into a prescribed space and location. So, you ask, how do I limit the size of the table and allow movement to the extra data?
Well, the first response is to ask if you are sure that's what you want. There are no scrollbars on paper, so if you try to print your form, you will not be able to see all of the table content. You may actually want to limit the table size to some maximum number of rows. It is very easy to make the insert trigger (button) become non-relevant when a maximum number of rows is reached. The easiest way is to add a
<xf:trigger ref="item[3 >= last()]"> <xf:label>Insert <xf:insert ev:event="DOMActivate" nodeset="../item" at="index('repeat-cart')+1" position="before"/></xf:trigger>
Relative to the post on Nov. 3, two things happened. The
However, when the total number of item elements is three or lower, then the trigger
Finally, let's assume you don't care about the printing problem and so you really do want to have a 10 row table fit into the space of three rows with scrolling. Well, an XForms repeat just repeats things, it doesn't decide whether or not to grow a scrollbar. But, you can use an XPath predicate on your XForms repeat to control how many nodes it is showing. In fact, we highly recommend that you do use such predicates to keep repeat tables down to a reasonable size for efficiency. There's no point in creating 200 rows when the user can only really see 10 or 20 at a time!
To do repeat scrolling, you should create an instance data variable that can hold an index of your current location within a larger dataset. You could have a second variable to control the size of the table, or you could just pick a constant. I will pick 10 here. You then adjust your repeat nodeset as follows:
<xf:repeat nodeset="item[position() >= instance('temps')/curpos and position() < instance('temps')/curpos + 10]"> ...</xf:repeat>
Now, you just need scroll up and scroll down triggers (buttons) that add or subtract 10 from the curpos variable using a setvalue action. Of course, you should also use the above trick to make those triggers become non-relevant when doing so would cause the curpos variable to go out of bounds (less than 1 for the scroll up trigger, and greater than count(item) for the scroll down trigger).[Read More]
John M. Boyer 060000VMNY 806 Visits
We usually think of a document as a file. A file is, at the operating system level, a unit of information. The OS can associate properties with it (like who can read it or write it), and the file can easily be moved from directory to directory or computer to computer with standard tools like FTP, email, or flash drive.
The XML recommendation defines a document as that which conforms to BNF rule #1 of the XML syntax. Of course it is useful if an XML document is the sole content of the unit of information being transferred, i.e. if the XML is the only thing in a file. This is because XML tools tend to be built with programming languages that expose services of the operating system.
However, the most important part of an XML document is capable of being contained within another document. In fact, XML canonicalization strips away the part of an XML document that would prevent the result from appearing as element content.
This is the idea that XForms uses when it expresses XML instance data within a larger document, and specifically within the
This paradigm works well enough if you think of a form as an ephemeral view of a back-end database or content repository. However, people find it easier to work with a document in the file sense of the word. we like to save our document, email it, reload it the next day or on a new computer, and especially on a computer that has become disconnected from the net or from the original web application context that would be associated with the ephemeral view kind of form.
Enter the document-centric approach of Workplace Forms, the technological basis of which is an XML document whose element names belong to the XFDL and XForms vocabularies. Moreover, the underlying XML data document is still incorporated by the
A key difference is that being a document implies that the form with eventually have to be re-serialized, e.g. to save or sign the form. XForms doesn't really talk about how to reintegrate the live, modified XML data back into the document, but in truth it's really not that hard. The main technical hiccup on the XForms side is that the
The conclusion, though, is that a Workplace Form, or XFDL document, is a cross between an office document and an mobile intelligent agent. It can be saved, reloaded, printed, emailed, archived, signed, submitted to web application and portal servers, and participate in document management, records management and content repository processes. This is how the notion of document makes a form be more than just an ephemeral view of data.[Read More]
John M. Boyer 060000VMNY 796 Visits
On a recent ad for a movie or TV program, I heard an irreverent but nonetheless funny line about there being no team in I. Good for a laugh, but it did remind me of that more deeply satisfying idiomatic saying: There is no 'I' in team. As well, it reminded me that I wanted to say a thing or two more about a recent interview I did with Jonathan Erickson for the Dr. Dobb's Journal Developer's Diary.
The purpose of the diary is to provide snapshots in the lives of real developers at various stages of their careers to find out what they do, what makes life harder for them, and what makes their job easier. I was honored to be selected, and Dr. Dobb's did a good job of relating what I said. If you read the print version, you'll see that they said 'Turing incomplete' instead of 'Turing complete' but that was immediately corrected in the online version. But the main thing I wanted to talk to you about was the question of what makes my job easier because I talked about two things, but they only had space for one.
In response to the question of what makes my life easier, I would say that I see my job as a senior technical leader being comprised of both challenging problems to work on as well as interesting people to work with. On the problem-solving side, I made mention of the important role that algorithm analysis has played in not only effective problem-solving, but in seeing where the problems really are and how to either solve or sidestep them as needed. However, I'd like to at least make public my thoughts on the people side of my role, too.
Specifically, I would like to call your attention to the Workplace Forms team and the XForms community. Whether I am directly participant in a collaboration or simply observing the members of either group, I have to say that the level of dedication you demonstrate inspires me. You, the people, certainly give meaning to my own work, and I am proud to be working with you to build Workplace Forms and XForms.[Read More]