Category format

Recipe XSD

I have been playing with one of the XSD-files of StructuredBlogging and hacked it to a simple-recipe XSD. You can find it here. I used the structure of a recipe as defined by Advenio. I still have comments on that structure, but then this is just a simple recipe.

Changing the file into a recipe was pretty straightforward. There are however several conventions, which can defined and used for future MicroContent definitions:

• title - I assume that each MicroContent item will have a title, which describes the item. Such a title can also be used as the title for the blog item, when a MicroContent item is published in a blog-container. This recipe however uses the element name name. This could be useful in order to distinguish it from the blog-item title. The simple-review xsd of sStructuredBlogging uses the element-name review-title. Clearly some standard approach is needed;
• Type specific elements - Some elements are specific to the MicroContent type. For instance the element ingredients are likely to be useful only in the context of a recipe. To make this point clear one could begin each element name with a ‘recipe-’ part. StructuredBlogging seems to do a similar thing;
• Tags, Rating - an element that gives a list of tags or keywords seems to be very popular nowadays and is thus relevant for any type of MicroContent. A standard element name would be useful. I used the element name keywords here. By the way the same is true for the element name rating;
• source - the recipe definition has an element source. This element indicates where the recipe was obtained. The source can be a person, a book, a website, etc. This reminds me of the product element of the simple-review definition. And I could replace the source element by this product element. It is another reason to define a product in a separate XSD. And then I could either use the product-author, product-link or product-name to define the source of the recipe;
• enumeration - I do not know how to encode enumerated types in XSD. So that part is postponed until I know how to do it;
• lowercase - I encoded all element names in lowercase and used dashes (”-”) where needed;

Categories/tags: format
PermaLink • Comments • TrackBacks

SB Envelope

As a second step I had a look at the Envelope that has been defined on StructuredBlogging. And piece of MicroContent should have this envelope around it. This piece of XSD is a bit more complex. A xml-structured-blog-entry contains a sequence with generator and any elements. I assume that the any element is the review or whatever. The generator element contains an id, a type and a version attribute. I do not know what these are supposed to mean. A unique id of a structured blogging item? The type of the Item (review, event)? And what is the version?

Anyway some comments:
- I have mixed feelings over the word structured-blog-entry here. I see MicroContent broader a a blog thing. There are also other ways to publish an item. Just call it SB-item.
- I miss the minOccurs and maxOccurs of generator. What happens when I publish multiple items (blog-entries)?
- I miss the dates and author. If we assume that an item structured-blog-entry can live on its own, it should contain all relevant information. Thus author and dates should be added here. I guess they assume that either the items in the RSS-container or the items of a blog in a HTML-container contain this information. This is however a big assumption and implies that SB-entries can not live outside these two specific containers. ALso the HTML-container for blog is ill-defined.

Categories/tags: format
PermaLink • Comments • TrackBacks

OPML to RDF

Danny Ayers has a nice entry on translating Attention OPML to Attention RDF. Translations are an important issue. As I got a bit enamoured with OPML, I had a closer look.

I noticed that the OPML he used from Attention.XML took a different approach from mine. Their approach is that a single outline tag corresponds to a single record. But that breaks down a bit with the followedLinks= attribute, which has two values.

I used OPML attribute field to define the namespace, OPML attribute values to define the namespace attribute and an OPML value as the value of the namespace attribute. Thus

. And each namespace attribute would have his own outline child. This looks a bit cleaner and extensible to me. And is OK for parsing as well.

Danny Ayers does not like syntax of the XHTML translation for the translation of the OPML-file. I agree here. I like the structured blogging approach better. And I like his RDF translation. It is much cleaner.

I agree with him that modelling and display are different subjects. Use OPML for modelling and start translating from there.

Categories/tags: format
PermaLink • Comments • TrackBacks

Attention

I did not yet comment on the ideas of attention.xml. The main reason was that I did (and still do not) see the issue. Danny Ayers weblog-entry reminded me to have a second look.

As I understand it, attention.xml is just about transforming a blogroll to nice XHTML. I tried on the attentiom.xml website. I uploaded my blogroll opml-file and got a nice website in return. Great, nice, so what.

Ted Leung enlightened me here. Attention is what I used to call user profile. In this profile is an extended history on what the user does online (pages he visits, books he buys, RSS feeds he reads, etc.). Great stuff for a marketeer. Or that was promised to us. Recommendation Services. I used to be an intrigued user of Firefly (remember?) and had a project with NetPerceptions on recommendations on bars and restaurants. And I still have a look at what Amazon recommends me. But all in all, I am a bit disappointed. Businesses still have trouble in using my blog(s) and recommending me nice things.  People have been dreaming on this subject. So don’t bother where my attention lies.

Categories/tags: format
PermaLink • Comments • TrackBacks

Flexibility of OPML

Marc Canter describes how flexible OPML can be. I guess I never encountered this flexibility due to the rigidness of my outliners. From his description I have the impression that any relational database can be put in an OPML-file. Sounds flexible enough to me. I like his word “structure editor” for an outliner by the way. Gives me the feeling that outliners have still a long way to go. Interesting.

I should learn XSLT by the way. Then we can be transform one OPML review namespace to another.

Categories/tags: format
PermaLink • Comments • TrackBacks

OPML attribute namespaces

One of the things that I encounter with translating MicroContent to OPML is the various namespaces. Where the namespaces come form must be visible somewhere in the OPML-file. I suggest the following:

Each MicroContent item is encoded a separate <outline> </outline> pair. This is the outer parent. All information relevant to an Item is encoded as children of this parent. This outer parent has two attributes. A text= which can be used to give human friendly explanation of the item, such as “book review”, “bookmark”, blog-entry”, “recipe”. The user is free here.

The second attribute depends on the namespace that has been used. For Alf Eaton’s reviews it could be RVW=, for a del.icio.us bookmark del.icio.us= (it seems that dots are allowed), for a Structured Blogging review it would be SB= and for the output of a recipe from the MacGourmet application it would be MacGourmet=.

And the user must be able to find the meaning and definition of the tags somewhere on the web. This could be encoded as a value of the second attribute. Thus RVW=“http://alf.hubmed.org/rvw”, del.icio.us=“http://del.icio.us/”, SB=“http://www.structuredblogging.org” and MacGourmet=“http://www.advenio.com/macgourmet/”.

Sound like a pretty lowercase approach to me. I do not think that OPML allows for varying attribute tags or is that just a limitation of my outliners? Variation makes it possible that multiple namespaces can be used in a single OPML-file, or better: one can mix multiple MicroContent types in a single OPML-file.

Categories/tags: format
PermaLink • Comments • TrackBacks

SB ideas

Thanks to Marc Canter I found the posting by Sébastien Paquet again. In the post he is basically dreaming of a review aggregator. In order to get such a service working, you need structured blog post (review posts).

He says that the interfacing for the creation of reviews should have specific fields, such as an ID, in order to make aggregation easier. And indeed something like this should be working for any review (dare I say MicroContent type). Next such a review could be posted to your blog, added to your RSS or submitted to a service.

That is one of the reasons that I like MicroContent clients. They offer structuring for a slightly different reason (management), but also allow for publishing and exporting in various formats.

Categories/tags: format
PermaLink • Comments • TrackBacks

OPML and MicroContent

Here an interesting article on OPML. I still wonder whether OPML can be adapted to support Any type on MicroContent. I guess the answer is yes, but some adaptations are necessary.

Categories/tags: format
PermaLink • Comments • TrackBacks

Microformats

XML.com’s Micah Dubinko discusses Microformats. I am not sure whether I understand it fully, but it seems a light way of going ahead with MicroContent. I wonder whether this is a too light way and something more fundamental is needed. I like the structure blogging idea better. But it is all about adoption: what works better?

Categories/tags: format
PermaLink • Comments • TrackBacks

HR-XML

I came across this Human Resources standard organisation. They also have a standard for Resumes, which might be interesting for MicroContent creators. I wonder whether there is, or should be, a link with FOAF. The XMLResume site provides some help to build your resume. I will have a go at it.

Categories/tags: format
PermaLink • Comments • TrackBacks

Container Standard

I am looking for a container XML standard. This standard should be able to contain multiple Items, which are defined in XML themselves. Ideally this standard would also allow from grouping Items into folders and subfolders. You might call this a List standard as well. The standard should also allow for the annotation of the entire list and each folder.

The only candidates that I am aware are RSS and OPML. But RSS does not really allow for XML-items, although only can put anything into an item-field. And RSS does not allow for folders. OPML is great for folders and one could put an XML-file into an attribute, but that is not really very nice.

Categories/tags: format
PermaLink • Comments • TrackBacks

OPML and RSS

Recently Adam Curry was again very happy with OPML. And in a sense he is right. It easy to open such an OPML-file and it works great. And it is even more amazing that it works as import and export for RSS-feeds. Who standardised those attributes (type, text, title, xmlUrl, description, version, htmlUrl)? I assume that Dave Winer wrote that down somewhere.

But why not create a blogroll standard? I guess the answer is that this works, so why bother.

Categories/tags: format
PermaLink • Comments • TrackBacks

Media RSS

With the coming of Yahoo Video Search, they also decided that a video needs metadata as well. Well that sounds like a good idea. They decided that RSS needs a few new subfields of the [cite]item-field[/cite]. So they created the fields , , , ,

. And each field has multiple attributes. I have no idea whether these are the good fields to have for a video. I have a feeling that it is a bit limited. I am not aware of any other attempts.

But why the combination with RSS? Video-content should be able to stand on its own legs. Create a standard XML-based mark-up for describing a Video in all its details. Eventually this should also contain time- and image-based mark-up. And then create something to embed such a XML-file into RSS. But that is only needed if you want to have a feed with multiple video’s in a single RSS-file. Otherwise just access the original XML-file.

But as Marc Canter says it is already applaudable that they think of metadata in the first place. OK.

Categories/tags: format
PermaLink • Comments • TrackBacks

GPS and EXIF

Any location-based piece of Microcontent should have metadata with location fields. Engadget reports on an experiment with GPS.

I wonder what the is the best way to record such metadata. EXIF is metadata embedded in a JPEG image file. I also read somewhere that Flickr allows to create separate metadata tags. I think I like the approach to store metadata in a human readible file and not in the image.

Categories/tags: format
PermaLink • Comments • TrackBacks

Microcontent containers

I saw on PaidContent a reference to a report by OCLC on Containers, i.e. books, CD’s, etc. And I thought immediately Microcontent. I like the word container for a book, a CD, etc. But this word can also be used for a weblog, an RSS-feed, a XML-file, etc. Most of these digital containers container similar Microcontent items. Thus one must distinguish between Microcontent and its container. An file is something different from the content. I wonder whether the Dublin Core makes this difference? Something to look into. Anyway Microcontent must be able to fit in a variety of containers.

As the report makes reference to Microcontent, I have downloaded it.

Categories/tags: format
PermaLink • Comments • TrackBacks