Moodle in English: Are there any standards for learning interaction data in XML?

Are there any standards for learning interaction data in XML?

by Matt Bury - Saturday, 28 November 2009, 11:32 AM

Hi,

I'm trying to work out a data model for Flash learning applications in Moodle. It's for learning interaction data so it's things like URLs to media files, questions, answers, options, instructions, etc. basically everything that a Flash application needs to display a learning interaction.

So my question is, what standards have already been established for this? Which are the most commonly used and the best supported? i.e. Is there already handy software available for generating this kind of XML code so that non-developers can easily generate learning content?

BTW, I intend to duplicate whatever structure I use in DB tables as far as possible so that the most commonly used learning interaction data sets can be handled via Flash Remoting as a faster, lighter alternative to slow, heavy XML.

Thanks in advance. Looking forward to your replies.

Matt

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Joseph Rézeau - Saturday, 28 November 2009, 6:12 PM

Hi Matt,
When you say " It's for learning interaction data..." this sounds ambiguous to me. Do you mean
a) it's for learning "interaction data" i.e. you are interested in learning all about "interaction data"
ob b) it's for "learning interaction" data i.e. you are interested in data concerned with "learning interaction"?
TIA
Joseph

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Frank Ralf - Saturday, 28 November 2009, 11:11 PM

Hi Matt,

There are some XML FAQ but I'm not quite sure I understand what you're after

Frank

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Sunday, 29 November 2009, 6:52 AM

Thanks for the replies!

@Frank, yes this is what I mean: http://docs.moodle.org/en/Moodle_XML - This is the XML for the Quiz activity module with questions and answers. In this case, it looks too basic, i.e. pure text and HTML , and no multimedia, for the kinds of things that Flash does.

From my general research into XML for Flash in an e-learning context, I haven't found any widely used standards - that's why I posted this thread. It looks like I'll have to come up with something useful and easy to implement.

The great thing about XML in Flash, is that it's very easy to handle (Flash supports E4X notation) and There's ways of writing code that means you can alter the XML structure to an extent and still maintain backward compatibility.

Of course, there's nothing stopping other developers coming up with generic XML formats for specialist applications and using it with the SWF Activity Module. I'd just like to have as many people using the same standards as possible to make Flash learning interactions interchangeable between different Flash learning apps. and different LMSs/CMSs. For example, loading the same XML file into different Flash learning applications to create different learning interactions.

Any advice on XML formats, caveats, best practices, etc. would be greatly appreciated.

Thanks again in advance!

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Penny Leach - Sunday, 29 November 2009, 4:21 PM

Hi Matt,

Have you looked at IMS QTI? It's the generally accepted standard, (as far as I can tell) for describing questions and quizzes.

Moodle does export its core questions types in QTI, but there's no way to get the feed of an entire quiz yet (a few custom implementations exist though).

If you're basing your work on Moodle 2.0, you may be able to get around this by getting Moodle's AMF implementation to export the QTI for each question as you need it.

Hope that helps

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Sunday, 29 November 2009, 11:09 PM

Thanks Penny.

I've looked at IMS QTI and I think from the perspective of loading it into a Flash app. parsing it and reading off data, it's too heavy. In fact, from the examples I've seen, the file sizes of around 175 - 200 KB would probably trigger the default 15 second time-out in a Flash app. Also, for the same reason that SCORM isn't very accessible to developers, QTI looks overly complicated and includes a lot of specifications that would be better handled by Flash apps.

Re: AMF, I fully intend to integrate the eventual data model with Flash Remoting - both AMFPHP 1.9 (Moodle 1.8 + 1.9) and Zend_Amf (Moodle 2.0). AMFPHP is much easier for Flash developers to get started with and once they've got to grips with it, they can move on to Zend_Amf - from a Flash developers perspective they're almost identical, only AMFPHP includes a service browser application which makes debugging and development particularly easy.

Since Flash is the de facto e-learning application platform, it makes sense to get Flash developers involved and make Moodle accessible to them, don't you think?

Mmm... how to explain what I have in mind...

For starters, I'd like to use a relatively simple XML format, which is what most Flash developers are used to handling, that can easily be translated into a database model (Flash Remoting) to make it faster, lighter and more flexible. I think this would also offer good import, export and authoring options. It'd be pretty easy to develop a desktop XML editor with drag and drop functionality so that non-developer course content designers/instructional designers can author learning interactions quickly, easily and efficiently. I predict that the resulting file sizes would be a small fraction of current SCORM packages (currently in tens or even hundreds of MBs per package) authored in the most popularly used software (i.e. Captivate, Camtasia and Articulate) and authoring time, on a per interaction basis, would be much faster.

I also intend to facilitate keeping to a mini MVC design pattern for the module so that data, media and Flash apps. are kept separate and brought together at runtime, a bit like the idea of having "learning interaction cartridges". The XML and media files can be loaded into Flash apps., which operate as "projectors" that present the learning interaction. Each app. would present the same learning interaction data cartridge in a different way giving a different learning interaction. I have proof of concept up and running on my website Demo course: http://matbury.com/moodle/course/view.php?id=9 (guest access) This version uses Flash Remoting only, no XML, but most of the learning interactions are using exactly the same learning interaction data, "Common Objects".

If Dmitry Pupinin's Flash question type project is successful, that would also be an area where the same concept could be used and that would also allow full integration with the Quiz module and all its existing "tried and tested" architecture.

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Frank Ralf - Sunday, 29 November 2009, 11:17 PM

Hi Matt,

I just stumbled across this tutorial while learning about XML and Flash:

"Using an XML file with Flash"
http://flash-creations.com/notes/dynamic_xml.php

Is this the kind of integration you're after?

Frank

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Monday, 30 November 2009, 11:35 PM

Hi Ralf,

Thanks for the link.

My question isn't about how to use XML with Flash - I'm already pretty competent with AS 3.0 and E4X. BTW, the examples on that site are legacy AS 2.0 (Flash Player 8 and before) and probably one of the more difficult ways of handling XML in Flash.

My question is really about how to organise the XML data tree to make as useful as possible for the majority of cases. However, I think that a one-size-fits-all approach probably wouldn't work and we may end up with a range of different standards for different subject areas (I work in EFL/ESL) so where chemistry and molecular biology might already have their own XML standards (I don't know), foreign languages don't appear to.

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Joseph Rézeau - Monday, 30 November 2009, 5:48 AM

Matt:

Since Flash is the de facto e-learning application platform,...

I'm really surprised at that statement. Can you reference it, please?
TIA,
Joseph

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Monday, 30 November 2009, 11:54 PM

Hi Joseph,

Gladly:

http://www.upsidelearning.com/blog/index.php/2009/06/15/using-silverlightexpression-blend-for-elearning-development/

http://www.elearningguild.com/pdf/2/111703dev-h.pdf

http://bdld.blogspot.com/2009/04/adobe-elearning-suite-cs4.html

http://michaelhanley.ie/elearningcurve/whats-the-difference-between-flash-and-captivate/2009/05/13/

http://www.commlabindia.com/elearning/rapid-elearning-development.php

http://elearningdevnews.wordpress.com/

Also a search for Flash on the e-learning guild's website returns around 100 results and in Google results for "Flash" + "e-learning" return 21,600,000, whereas "Moodle" returns 2,140,000.

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by sam marshall - Monday, 30 November 2009, 9:37 PM

Hi,

I think you are asking for the impossible here - you're looking for a 'standard' format, which in this area basically means IMS, but you're also looking for it to be practical and useful. That doesn't happen, in my opinion, which is why we have a lot of IMS standards that add very little value behind 'woohoo, it's a zipfile with some HTML in' - and a great deal of complexity in getting to that point.

Yes, I'm pretty down on IMS standards. Sorry. That's just a personal opinion, certainly not one my employers would agree with. We do support some of them in code I'm involved with.

If you want the job done well, it's probably easier to do it yourself - if you define something really simple and consider how it's implemented so that other platforms could probably handle it (for example, make sure it could be implemented by other VLEs and in other client platforms apart from Flash)

--sam

PS While I'm throwing around personal opinions, I'll be voting for (X)HTML+JS+CSS as the 'de facto e-learning application platform' - and for Flash to be 'first up against the wall when the revolution comes'

Average of ratings:Useful (1)

Re: Are there any standards for learning interaction data in XML?

by Martín Langhoff - Monday, 30 November 2009, 10:56 PM

> I'll be voting for (X)HTML+JS+CSS as the 'de facto e-learning application platform'

On that track, Bryan Berry and company (from the Nepal OLPC deployment) are working on this track. The title is a bit over the top, but the gist is that he agrees with you, and is hacking on it:

HTML5 + JS: The Future of Open Education
http://www.olpcnews.com/content/education/html5_js_future_open_education.html

Average of ratings:Useful (1)

Re: Are there any standards for learning interaction data in XML?

by Mauno Korpelainen - Tuesday, 1 December 2009, 5:03 PM

Martin,

thanks for the nice link...and all these well-grounded opinions that you and Sam have given about performance and future of web (apps) wink

I recommend also this excellent video by Brad Neuberg which collects most of the good new features of HTML5, future javascript etc to everybody:

or higher quality version on Vimeo:

The only thing that slows down a little the introduction of HTML5 ( and totally blocks XHTML5 ) for larger audience is IE and like Sam noticed 30-40% of browsers in Africa and Asia are still today using IE6.

Flash is not yet dead but ... three cheers for the winner, HTML5! big grin

Today

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Wednesday, 2 December 2009, 4:28 AM

Hi Mauno,

That's shocking! AFAIK, IE6 doesn't support Flash very well and still has the annoying and problematic "Click to activate" EOLA. Good to see Firefox is doing well there though. That'll make a lot of web developers happy

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Mauno Korpelainen - Wednesday, 2 December 2009, 5:47 AM

Well...surprisingly in China 95.5% of browsers in today's Stat Counter are IE and 65.9% of them use IE6 (and modified Windows XP)

But the situation will change rather quickly - in Europe IE has already lost the markets (total 47.3%) and for example here in Finland the most popular browser in today's stats is Firefox 3.5 (30.6%), second Firefox 3.0 (18.8%), then we have IE versions 7+8+6 (15.4+15.2+6.2% = total only 36.8%) followed by Safari 4 (4.3%), Chrome 3 (3.1%), Opera 10 (2.5%) and other browsers 3.8%

I really wish that all countries around the world could follow the trend and start using modern browsers Firefox, Safari, Chrome and Opera - IE has some years time to follow web trends and add new features - and at least current web standards - to IE9 but Microsoft may lose also the global game soon permanently on browser markets.

Average of ratings:Useful (1)

Re: Are there any standards for learning interaction data in XML?

by Marc Grober - Wednesday, 2 December 2009, 6:53 AM

Apples need to sell hardware I think argues against it providing a full OS X version for pure pc hardware, while hackintosh usage is minimal and pc vendor sales with linux are modest. But I think is all it would take is for a handful of hardware vendors to move from windows and windows is history- should a large government contract go for pc's w/o windows and things could easily start sliding for Redmond

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Mauno Korpelainen - Wednesday, 2 December 2009, 7:59 AM

Marc,

personally I believe that Windows has a much longer expectation of life than IE - if nobody is using IE after some years Microsoft will give up and they make a good deal with Apple (integrate Safari to Windows) or buy Opera...

Or desktop PC:s with Windows might simply get replaced by some new technology - in fact Microsoft has many nice visions http://www.officelabs.com/pages/Envisioning.aspx

but I don't see any IE in those vision movies... wink

And it is also possible that IE9 or IE10 will suddenly have V8, TraceMonkey, SquirrelFish or Futhark inside (with big money).

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Mauno Korpelainen - Wednesday, 2 December 2009, 9:27 AM

One correction - IE9 will use Direct2D to improve its rendering performance and Firefox have stated their intent to do the same - and sooner! And Google will follow... Are they really starting to play "war games" with graphics cards - WoW

http://en.wikipedia.org/wiki/Direct2D

Edit: Firefox 3.7 alpha alredy had full Direct2D rendering 2 days ago - for the first time a web browser can render a website using the power of a PC's GPU rather than its CPU - a move which can dramatically speed up loading times. But also Flash 10.1 brings GPU acceleration which means your system will get extra assistance when playing streaming videos online...

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by sam marshall - Wednesday, 2 December 2009, 6:45 PM

In terms of JavaScript performance, benchmarks on IE9 beta show it's on par with Firefox (which is slower than Safari/Chrome but not bad); they've integrated their own JIT compiler.

IE is a bit behind but I think IE8 is already not a bad platform for delivering webapps (performance aside) and IE9 should solve most of the performance issue. The problem is very slow upgrades - i.e. we have a high percentage of IE users still on 7 and even 6, because updates either aren't available for older OS versions (in the case of 6), or are not automatically pushed. This compares to the profile of, say, Firefox where nearly all users are on a recent version because updates are automatically pushed - basically nobody is still using 2.x). There may be similar problem with Safari though because it's also linked to the (Mac) OS.

So in other words I don't think IE9 will hold back web development too much, what will hold it back is that (if current trends continue) upgrade to IE9 will be slow.

--sam

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Mauno Korpelainen - Wednesday, 2 December 2009, 9:21 PM

It's not necessarely all true...

Developers of MS have used this image

graph

to show how well IE9 is running but it was drawn before FF3.7 alpha tested Direct2D and if the same tests were done with latest test versions IE9 is again slower than Webkit browsers and Firefox. The benefit of using Direct2D rendering depends on used graphics card, used OS and code of test page...

And speed is not the only problem... http://a.deveria.com/caniuse/ (Compatibility tables for features in HTML5, CSS3, SVG and other upcoming web technologies)

However - http://www.webmonkey.com/reference/Internet_Explorer_9_might_support_SVG_and_CSS3_transformation_even_canvas

IE has never had native support for SVG or Canvas - they prefer VML and the favourite song in MS seems to be "I did it my way..."
IE needs always workarounds - for example Adobe SVG plugin that is no more supported or SVGWeb ( http://code.google.com/p/svgweb/ ) that "brings SVG support" to IE... about 8 years after SVG became W3C recommendation... no - not really - SVGweb just translates SVG to Flash and users of IE need both extra SVGWeb script and Flash plugin...

Google gave also workaround for Canvas with http://code.google.com/p/explorercanvas/ but it does not make IE9 any better...

In addition to those difficulties in upgrading IE6 you mentioned many versions of Windows XP with IE6 may also be pirate versions or test versions...

IE9 will be officially released in 2011 if old tradition continues so upgrading IE6:s and IE7:s to IE9 may be slow also for that reason - but then only few people will be using IE - if the current trend continues... wink

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by sam marshall - Wednesday, 2 December 2009, 10:57 PM

1) Isn't that graph one of the JavaScript tests, which mainly test language performance rather than rendering? If so, DirectX acceleration shouldn't do much.

2) I think it's more useful to compare with current browsers. Current Firefox is fast enough, if 3.7 is going to be faster that's nice, but whatever. We need IE to get to that 'fast enough' state. IE8's not so bad, 7 really creaks...

I think if IE developers were lying they'd probably come up with a graph that didn't show them as still the slowest available browser.

I agree IE currently does not support lots of features, as well as being slow. However I don't think that is the problem. IE8 supports plenty of features - sure, canvas provides opportunities to make applications totally inaccessible develop new kinds of application, but it's not really essential. Ditto for SVG - hey, at least PNG works now. Also, IE8 does sort of support the local storage from HTML5. Ish.

Anyhow, basically my point is, there is a certain baseline 'rendering doesn't suck, performance is okay, basic (if not advanced) features work' which currently IE8 meets apart from the 'performance' bit, and IE9 should meet. But there's a real problem with all those older versions hanging around, and the problem will be perpetuated if (as they did with Windows 2000) Microsoft decide that, say, IE9 won't run on XP. By this stage that would be a kind of reasonable thing for them to do, but it would certainly hold back updates.

--sam

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Mauno Korpelainen - Thursday, 3 December 2009, 6:41 PM

Sam,

wasn't some rumor/legend saying that 640K ought to be enough for anybody... wink

I agree. The main benefit goes to Microsoft and IE - current Direct2D is built on top of Direct3D, and offers some compelling performance and visual quality improvements over GDI and GDI+ but it can be used only on Windows 7, Vista and Server 2008 - it will not help at all XP, Linux or Mac users. From Microsoft's point of view the optimal solution requires people to ugrade old XPs and buy new versions of Windows - Mainstream Support for XP ended on April 14, 2009 and Extended Support will end on April 08, 2014. Microsoft is now finally running a world wide campaign through June 2010 that'll urge users to upgrade IE6.

Direct2D was introduced about year ago with some clear goals: http://blogs.technet.com/thomasolsen/archive/2008/10/29/introducing-the-microsoft-direct2d-api.aspx

Direct2D is designed primarily for use by the following classes of developers:

    * Developers of large, enterprise-scale, native applications.
    * Developers who create control toolkits and libraries for consumption by downstream developers.
    * Developers who require server-side rendering of 2-D graphics.
    * Developers who use Direct3D graphics and need simple, high-performance 2-D and text rendering for menus, user-interface (UI) elements, and Heads-up Displays (HUDs).

http://msdn.microsoft.com/en-us/library/dd370990%28VS.85%29.aspx

Together with DirectWrite (support of it is also implemented to FF) users of Windows 7, Vista and Server 2008 will get also high-quality and soft text rendering:

Today's applications must support high-quality text rendering, resolution-independent outline fonts, and full Unicode text and layout support. DirectWrite, a new DirectX API, provides these features and more:

    * A device-independent text layout system that improves text readability in documents and in UI.
    * High-quality, sub-pixel, ClearType text rendering that can use GDI, Direct2D, or application-specific rendering technology.
    * Hardware-accelerated text, when used with Direct2D.
    * Support for multi-format text.
    * Support for the advanced typography features of OpenType fonts.
    * Support for the layout and rendering of text in all supported languages.
    * GDI-compatible layout and rendering.

The API supports measuring, drawing, and hit-testing of multi-format text. DirectWrite handles text in all supported languages for global and localized applications, building on the key language infrastructure found in Windows 7. DirectWrite also provides a low-level glyph rendering API for developers who want to perform their own layout and Unicode-to-glyph processing.

http://msdn.microsoft.com/en-us/library/dd368038%28VS.85%29.aspx

For normal web pages the difference may be small but for such pages that use a lot of graphics, animation or advanced features of text rendering the result can be rendered 10 to 100 times faster with soft, high-quality moves. On the other hand if Microsoft, Apple, Adobe and Google start creating browsers that are tied to certain versions of graphics gards or operation systems it's not at all good... and if Firefox can keep doors open to all directions it will be the number one browser forever...

I am a Windows user - I have always been - and I have tried to use Flash for some years (bought even Adobe Flash CS3 Professional) but no matter what Microsoft, Adobe, Apple and Google do the main power of coming decates should be in Open source and HTML5 that everybody can use and even if some apps were 10ms slower with javascript than with Flash (now) the most important thing that developers can do is to make such selections that help in developing easy to use and functional, free, high-performance web apps for all the people. Like those that we will be able to create with HTML5, CSS3 and javascript if Microsoft decides to come along...

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Thursday, 3 December 2009, 9:44 PM

Hi Mauno,

I think there's one major flaw in waiting for IE6 users to upgrade to later versions of IE. MS require users to go through a Windows authentication process in order to upgrade IE. I read in the Economist that most Windows OS's around the world are pirate versions and apparently in Thailand 95% of commercial and business software is pirated. I think the truth is that, whatever MS come up with, it won't be accessible to the majority of users in the developing world because they can't or won't afford what MS charges.

I think the most productive campaign would be to get the world to abandon IE altogether and let the likes of Chrome, Firefox, Safari, etc. take over (although I don't like the idea of Firefox and Chrome extending Google's dominance of the web). I don't understand why MS feel that they have to develop their own browser and force it on everyone in the first place. The Internet, without IE, would be a nicer place to work, for one thing. Just take a look at the source code for any major site and look for the list of "if" conditions relating to IE!

Do you think Moodle should join this Norwegian campaign?: http://www.wired.com/epicenter/2009/02/norwegian-websi/

The article says that even MS are joining in and urging users to upgrade to IE7 or later. It would be an ideal opportunity to encourage users to switch to browsers other than IE.

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Mauno Korpelainen - Thursday, 3 December 2009, 10:17 PM

Not necessarely war - Google and MS are fighting already...

European Commission was using quite heavy methods in pressing Microsoft to create different version of Windows 7 for Europe but the result was

http://www.engadget.com/2009/10/07/european-commission-likes-windows-7-browser-ballot-screen-will/

Maybe we should tell all our neighbours that they don't need to use IE - there are 4 better browsers available. The only thing that strikes back in moodle is editor (HTMLArea) that was written to support only IE and Firefox but that will be fixed in moodle 2.0 with TinyMCE.

2³² is already 4 294 967 296 so if everybody finds 2 new neighbours who have not heard this message...after 32 message rounds quite many people might consider changing their browser...

Microsoft should just follow the same "Web standards" as other modern browsers and most people would be happy. They don't want to use for example SVG or Canvas because these tools make VML useless - or any other changes that could drop their leading position on desktop. It's like selling competing products. wink

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by sam marshall - Thursday, 3 December 2009, 11:45 PM

I agree some of the HTML5 features are nice, but it isn't really a 'web standard' yet. Microsoft have finally done a good job in supporting web standards (HTML 4.whatever, CSS 2.whatever) in IE8 and should be applauded for that. It's not yet time to condemn them for not supporting HTML5...

In practice IE exists and will continue to be a major force for some time.

--sam

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Mauno Korpelainen - Friday, 4 December 2009, 1:52 AM

Sam,

I can't find any text where I had said that HTML5 is a Web Standard. In previous comment I said: Microsoft should just follow the same "Web standards" as other modern browsers and most people would be happy.

I could aplaude to Microsoft for a good trial and basicly IE8 is ok - but go to http://acid3.acidtests.org/ and test how well your IE can handle ACID3 test page. Acid3 is a test page from the Web Standards Project that checks how well a web browser follows certain selected elements from web standards, especially relating to the Document Object Model and JavaScript. Microsoft itself has continued to downplay the importance of passing the Acid3 test. Current builds of IE 9 score 32/100, up from 20/100 for IE 8 and 14/100 for IE 7.

Internet Explorer has never supported the application/xhtml+xml MIME type, which essentially means not supporting true XHTML as specified by the W3C
XHTML 5 is the XML serialisation of HTML 5 and because XHTML5 must be served with an XML MIME type, such as application/xml or application/xhtml+xml Internet Explorer can't render any XHTML5 pages.

+ Internet Explorer has never supported SVG which has been Web Standard since 2001 (8 years)

In practice IE exists and will continue to be a major ~~force~~ pain for some time.

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Frank Ralf - Friday, 4 December 2009, 2:00 AM

Just stumbled across Processing.js:

"Processing.js is an open programming language for people who want to program images, animation, and interactions for the web without using Flash or Java applets. Processing.js uses JavaScript to draw shapes and manipulate images on the HTML5 Canvas element. The code is light-weight, simple to learn and makes an ideal tool for visualizing data, creating user-interfaces and developing web-based games."

Looks quite promising.

Frank

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Mauno Korpelainen - Friday, 4 December 2009, 3:07 AM

Frank,

thanks for the link - yes, I have seen it and it looks promising but you did not paste the next two paragraphs of introduction:

"Processing.js is explicitly developed for browsers that support the HTML5 <Canvas> element. Processing.js runs in FireFox, Safari, Opera and Chrome but will not be supported in Internet Explorer until Microsoft catch up with ISSUE 15.

Implementing Processing.js in Flash or Silverlight is not recommended as Java already occupies the browser-plug-in space for this library. For users wishing to run Processing.js in Silverlight, see Paul Irish's Silverlight implementation. Using Explorer Canvas with Processing.js typically results in unusable frame-rates for moderately complex visualizations."

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Mauno Korpelainen - Thursday, 3 December 2009, 10:45 PM

And yes - we should have some way to get rid of those "if IE {...} tags"

Supporting IE6, 7, 8 and 9 in themes for different reasons might be worse than a nightmare - but all theme designers could drop IE6 support right away from their custom themes. 8 years extra work for IE6 should be enough.

In official version of moodle there may not be that option to drop even IE6 as long as so many people still use it.

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Ludo (Marc Alier) - Wednesday, 2 December 2009, 4:32 PM

Hi Martin,
thanks for the link to the Karma project. A bunch of my students are building an HTML 5 + js application to create and play JClic contents. So we will pay attebntion to their work.
Saludos
L.

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Monday, 30 November 2009, 11:58 PM

Hi Sam,

Just out of curiosity, what would you prefer to use in place of Flash? MS Silverlight? Java? Actually,what I've seen of HTML 5 is pretty impressive. What do you think?

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by sam marshall - Tuesday, 1 December 2009, 2:19 AM

Silverlight? Wow, you had to pick practically the only thing I like less than Flash...

(Surely on purpose there.)

Sentimentally, as I still think I'm a Java developer (despite mostly writing PHP), I'd prefer Java. However from a practical point of view, I think deployment of Java can be complex - partly because it tries to do a better job than Flash in fitting in with platform norms, which can make it more challenging to debug cross-platform, partly because until very recently it didn't include support for things such as media playback which are trivial in Flash, partly because historically it's been less reliable through version changes, and partly because the new JavaFX stuff which is supposed to compete with Flash, while certainly very cool, seems to compete with Flash largely by being rather like Flash (and less widely supported, and less reliable). Don't get me wrong, conceptually it's in theory somewhat better, but I sort of have a feeling it might join the long list of eighteen gazillion Java-based buzzwords/acronyms that nobody really needs to know.

At present, I like Flash for audio/video playback (because it works) and for animations (because that's what it was originally designed for and does an okay job at). I like HTML / JavaScript for everything else, with Java used only for complex applications that can justify being treated as a full application with testing and everything (probably the kind of stuff that launches from WebStart in a separate window).

If and when HTML5 is fully adopted and reliable, particularly the audio/video tags, I can probably ditch the audio/video Flash from those preferences! Not holding my breath though - we still have a small proportion of users on IE6 and a majority on IE7, so even if IE9 gained support for video tag etc, it'd be a Long Time Coming. Too far ahead to really think about.

All this is my personal opinion. I might think it sucks but we do use Flash a lot here, and some of it I've helped to integrate. For instance we built a 'flash (and java) library' feature into the system that publishes content, so that people outside the direct area I work in can create new, generic interactive features using Flash (or Java if required). For instance somebody built a Flash picture viewer with large zoom capabilities for arts courses so you can look at images in detail, and we integrated a nice open-source molecule viewer (Java applet this time) that lets you spin 3D molecule pictures around (not sure why, but it's entertaining*).

There are definitely some nice uses for Flash and Java but I think providing a 'shell' around content is not a good one - i.e. I think Flash/Java applets, if required, should be little interactive thingies within HTML pages where most of the content lives as proper text and images. Basically they should stay in their place and not get uppity and try to take over the whole website. That's it.

--sam

* I do not know very much about chemistry, or did you guess that?

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Wednesday, 2 December 2009, 4:56 AM

Hi Sam,

If you're used to using AS 2.0, I can understand your frustration with it compared to Java.

Funnily enough, ActionScript has been moving in the direction of Java for a while now. AS 3.0 is fully object oriented (not compatible with AS 2.0 and runs on a different Flash Player engine - AVM2) so you can essentially write full software applications that run in the browser. It fully supports XML, web services, JSON and it's fast enough to handle full 3D environments and objects, although not quite as well as desktop games. It also handles multimedia even better and Flash's AMF, which has open source support, is many times faster than JSON.

Visit my Splash page and play with my 3D Flash cube: http://matbury.com/ - This one's only pseudo 3D but I've been using 3D AS 3.0 libraries which run very smoothly. For example, see PaperVision3D: http://www.papervisionshowcase.com/

As far as I can see, using a run time compiled language platform such as JavaScript is always going to have performance issues. Currently, Google are pushing JS to its limits with the cloud. My PC is pretty fast and even using Google Chrome with its super fast JS engine, Google Wave presents a challenge for it.If you want to do something really snazzy, you have to embed Flash in it!

Just my €0,02!

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Mauno Korpelainen - Wednesday, 2 December 2009, 6:17 AM

Matt,

I don't think that nobody has said that you could not create nice applications with Flash and Action script 3.0...

Sam is a developer and I am more or less an "end-user" of different scripts but as far as I can see your last paragraph

"...using a run time compiled language platform such as JavaScript is always going to have performance issues. Currently, Google are pushing JS to its limits with the cloud. My PC is pretty fast and even using Google Chrome with its super fast JS engine, Google Wave presents a challenge for it.If you want to do something really snazzy, you have to embed Flash in it!"

looks strange and exaggerating big grin

Check that video link about HTML5 - Google has just started the optimizing of their Javascripts and even if those big applications may have much more code than your flash applications Javascripts are getting faster and faster all the time thanks to faster browsers and better written code & nice co-operation with HTML5, SVG, videos etc.

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Wednesday, 2 December 2009, 11:47 PM

Hi Mauno,

Perhaps you haven't seen this: http://gotoandlearn.com/play?id=105

Or this: http://theflashblog.com/?p=307

Flash Player 10 already has a rudimentary 3D graphics engine built in (most coders are still relying on 3rd party libraries for 3D that support stuff like Collada - standardised 3D XML files) and I think by the time FP 11 comes out it will have matured into something more useful. Also, FP 10 includes inverse kinematics, that's things that have hinges and joints to you and me.

Adobe have been quietly pushing the idea of enabling P2P communication in FP. Since freely accessible P2P would severely cut into their Flash Media Server products, I think they're talking about making it only work with FMS - Which kind of defeats the object but would bring down FMS bandwidth costs.

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Mauno Korpelainen - Wednesday, 2 December 2009, 7:18 AM

http://www.macworld.com/article/143691/2009/11/jscript.html

Of course different people have different opinions...

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Martín Langhoff - Monday, 30 November 2009, 10:49 PM

I predict that the resulting file sizes would be a small fraction of current SCORM packages (currently in tens or even hundreds of MBs per package)

Hi Matt,

you are a bit confused! SCORM packages are not very large because of SCORM. They are large because people add video, large images, etc.

Nothing to do with SCORM as a format. I've worked on some fantastic SCORM content that had small images, sounds loops and some creative programming. They looked great, played great, downloaded fast.

It's all in how it's built.

I agree with Penny, QTI is a good reference. The data model of SCORM is also good. You are right in that they look enormously heavy when you read the specs. I am often appalled with those specs too.

Two suggestions, however:

Skip the specs and find the examples. Simple usage of SCORM datamodel and QTI is often simple. There are a zillion options that different vendors wanted to get in there before they agreed to the spec. Just implement and use the subset you care about. Maybe later you find that some of those obscure options were a good idea to implement too, once you find the use case.
If the resulting XML is too large and verbose even with gzip compression (gzip is dictionary based, so it deals well with verbose XML stuff, but some verbosity is way too verbose), do a search-and-replace trick and implement it in shorthand. Or in a JSON-encoded variant.

The gist of the story is: those datamodels are a bit baroque, but their cores make sense. Following them is a good idea.

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Penny Leach - Monday, 30 November 2009, 11:36 PM

We do a lot of work here at Liip with a Flash presentation layer on top of the Moodle quiz and we do it with QTI and it works fine, performs fine, and and our Flash coders do great (although to be honest I don't think they like IMS much either ;) )

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Tuesday, 1 December 2009, 12:10 AM

Hi Penny,

Do you mean you're loading output from the quiz module into Flash or embedding Flash in Quiz pages?

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Penny Leach - Tuesday, 1 December 2009, 1:59 AM

I mean loading output from the quiz/questions via QTI into flash, yup.

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Tuesday, 1 December 2009, 12:08 AM

Hi Martin,

Have you got any examples you could point me to? I mean an examples of an XML file containing questions, options and answers for a specific learning interaction.

BTW, Flash has its own version of web communication, Flash Remoting (AMF) which is many times faster and lighter than JSON. It's the same format that it uses for video and web conferencing via media servers. My intention is to start out with XML but give developers the option of converting this XML into a database + Flash Remoting model.

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Tuesday, 1 December 2009, 12:26 AM

BTW, many thanks to everyone who has contributed. It's all been very helpful! =)))

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Frank Ralf - Tuesday, 1 December 2009, 1:04 AM

Hi Martín,

Your comment piqued my interest in SCORM (again) so I did a quick Google search for some tutorials. I added some links unceremoniously to the "See also" section of the SCORM FAQ so I find them again when I have the time to actually read them

Frank

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Wednesday, 2 December 2009, 4:18 AM

Hi Martin,

I agree that the size of SCORM packages is not due to the SCORM format, it's because of authors.

I think the majority of instructional designers and course content developers generally (I'm generalising here!) don't consider media file optimisation and don't regard file size as important unless they get complaints about download times, in which case, I don't think many of them know what to do about it. It's not the fault of SCORM per say but the authoring tools that designers use to create SCORM packages.

It is a problem if the majority of SCORM packages turn out to be in 10's or 100's of MB on average, especially if the learning resources are deployed on the Internet for users to download. This can result in excruciatingly long download times where users have ample time to go and have a coffee break or snack while they wait. Quite a usability issue if you ask me. This certainly appears to be the case.

I've also noticed that many SCORM authoring tools create Flash based quizzes that by my reckoning should be only a few KBs but publish to 100's of KBs. My most complicated Flash apps. are under 50 KB and they do far more than simple multiple choice and gapfill quizzes. I have no idea how anyone could regard this as acceptable.

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Michael Penney - Wednesday, 2 December 2009, 7:32 AM

Hi Matt, I think IMS LTI is (one day) the droid you are looking for

.

http://www.imsglobal.org/toolsinteroperability2.cfm

It's not there yet, but someday it might be a good framework - at least they have some good process flow diagrams

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Matt Bury - Thursday, 3 December 2009, 12:03 AM

Thanks Michael, that looks like just the droid I'm looking for.

May the force be with you. ;)

Average of ratings: -

Re: Are there any standards for learning interaction data in XML?

by Frank Ralf - Tuesday, 15 December 2009, 12:20 AM

Just stumbled across Talking to Drupal from Flash AS2 in the Drupal in Education user group. Sounds at least a bit related - at least I hope so

Frank

Average of ratings: -