MarkItUp - Thinking Products...

Subtext 1.0 "Nautilus" Lives

MarkItUp — Saturday, March 04, 2006 2:19 AM

Late last year when BlogML was a shiny new object I started to look around the blogging universe for potential blogging engines to target. One of the people that I contacted was Phil Haack who I knew was involved in an open source blog called SubText. SubText is a branch of the the old .Text codebase. This all happened in early October last year.

This morning when I synch'd up my aggregator I was very pleased to read this announcement from Phil:

http://haacked.com/archive/2006/03/02/Subtext1.0NautilusEditionReleased.aspx

I get a great feeling when I see projects like SubText come to life; it keeps me believing in the dream that I had when I started coding.

Anyways, because these guys have implemented BlogML, all of you suckers who have installed SUB can go to your Administration page and dump out your SUB blog into BlogML and then import it into "Nautilus"!

Well done guys, enjoy the glory!

Favour to ask...

MarkItUp — Friday, March 03, 2006 2:32 AM

Up until the weekend I had really only heard about Feedburner but hadn't taken the time to go and see what it's all about but after seeing this post I got interested in enough to move my RSS feed over. So now I have a favour to ask. If you are subscribed to my feed could you please change the URL that you use to my new Feedburner address:

http://feeds.feedburner.com/Markitup

This will help me to better understand how many people are subscribed to my blog and will also hopefully help me to ultimately produce better content.

Thanks!

Interesting data about the size of the blogsphere

MarkItUp — Wednesday, February 08, 2006 11:43 AM

Blogosphere Growth

BlogML - progress

MarkItUp — Saturday, August 27, 2005 2:32 AM

This morning I purchased the BlogML.com domain so that I will have a central place to publish the schema files and so that I can expose other tools from that site - such as an online validator to validate against the BlogML format.

I'm hoping to have the schema files and the other artifacts that have been created to support BlogML posted somewhere that they can be collaborated on until we have a final spec. Once we have the spec. nailed then I can publish it on the BlogML.com domain and create other cool things such as online validators etc. These files should be available by the end of the weekend.

This morning I also tested fully integrated BlogML import/export functionality into the admin console for SingleUserBlog. I was able to do an export to BlogML then delete all of my posts and comments and then restore (import) from the saved file. This was very cool. Because SingleUserBlog supports multiple data providers I was also able to export my file based content, switch providers and then import it all in to a database. This shows how cool BlogML will be for scenarios such as either migrating content or moving blog content between environments - such as moving from staging environment to a production environment.

Here is a link to the file that I imported and exported on my test system this morning so that you can see the format of BlogML: http://markitup.com/Data/Images/SUB.xml

Blogging API - some concrete ideas

MarkItUp — Friday, August 12, 2005 2:03 PM

A couple of people have asked me about the Blogging API ideas that I wrote about a little while back. Unfortunately I haven't really had time to write about this any further but I do have some concrete ideas about what needs to happen. Here's a very quick version of my "utopia" for this:

Basically, I think that each blogging engine should expose methods that allow you to export or import the entire content for the blog in a common format. Let's call it BlogML. An export for a blog could obviously be a huge file - but that's fine, storage is cheap and text compresses very well.

I've got a concrete sample of the schema for the format defined already and hope to post it in the coming days and perhaps it might promote some discussion and then gain acceptance somehow.

Think about this for a moment and what it means. Having a common format for blog content means that it would become very simple to import and export a blog; it would also be easy to backup and restore.

As an example, a feature such as this would allow you to go to the admin section for SingleUserBlog and choose an "Export" function" that would persist all of your content - posts, categories, feedback, trackbacks, etc - to an xml data file. Potentially, you could then take that file to another blog and import it, thereby making it simple to migrate content between applications or even between versions of blogging engines. These are things that have proven to be problematic in the past - for example, moving between early and late versions of .Text and then again, moving between .Text and Community Server.

I'll try to blog the schema definition that I have in the next few days and then implement it into SingleUserBlog as a proof of concept.

Finally, in my original posts I had discussed the idea of creating a common API. After sitting down and thinking about it for a while it became obvious that simply having a common API would not solve most common problems - such as interoperability. Having a common schema defined will solve problems such as these while freeing blog implementors to expose their API's as they choose.

Blog API Part 2

MarkItUp — Thursday, July 28, 2005 6:36 AM

I noticed that Chris has thrown down some of his thoughts about a blogging API. I'd really like to hear what some of the other "players" in the blogging industry think about this and who has been giving it some thought.

From some initial thinking I think that the API would need to be:

SOAP based. All of the major frameworks have SOAP API's these days.
Allow you to discover and obtain information about: Posts (including any attachments), Comments (including trackbacks), Categories.
Able to get information at a very granular level - ie: be able to return items by a variety of parameters and be able to perform queries which return both verbose and terse resultsets.

The benefit of being able to perform queries that perform terse results is that it would make it easier to perform atomic operations (allowing for re-tries) over an unreliable transporation medium.

So, looking at the shortcomings of the existing API's I feel that they make it very difficult to perform operations such as:

Get a count of posts
Get a range of posts
Cycle through all posts
Retrieve attachments
Discover linked documents
Query for Trackbacks or Comments (ie: types of feedback)

The existing API's seem to make it easy to post new entries or to edit existing content - provided that you already know the address of it. In other words, those API's do not lend themselves to auto discovery of content; and that's exactly what you need if you are going to go about writing a migration tool.

One thing that I should mention here is that, by talking about a blogging API I'm not talking about Rss. Rss and blogging have managed to evolve separately and should remain separate.

Blog API's

MarkItUp — Wednesday, July 27, 2005 5:42 AM

The request came out recently for me to build in a migration wizard which would allow for users to easily migrate content from their existing blogs into SingleUserBlog. This is a great idea and involves a problem which has come up quite a bit of late, that is... how to migrate content from one blog into another. We've seen this repeatedly within the .NET world as people have struggled to move content between .Text versions and then again when they tried to migrate into CommunityServer; and I'm sure that we'll see it again when CommunityServer has its next major revision. With the blogging universe having exploded I'm totally amazed that we haven't sat down to implement a common API for discovering and working with blogs - even more so given that we all see to agree on the virtues of having common API's and common Type systems.

The MetaWeblogAPI

Having heard of something called the MetaWeblogAPI [1] - which I knew was supposed to provide some sort of API based layer around blogs - I initially thought that building a single content migration tool would be pretty simple:

Locate the blog
Discover the category list
Import the categories
Discover the post list
Import the posts

Of course, importing a post would also involve the discovery of any attachments and metadata associated with it such as categorization and feedback entries. Once I started looking into the MetaWeblogAPI though it was apparent that this was a very primitive and hard to work with API compared to what I would expect to be dealing with. The API itself is based on XML-RPC which doesn't have the support that SOAP has within the .NET Framework. To work with XML-RPC I'd have to locate a third-party library and use it. The API is also very poor in that it has a very limited number of methods and the choices for working with data are virtually non-existant. For example, to get a list of posts I'd basically have to do a "GetRecentEntries" call and specify a huge number so that they would all get returned (hopefully); also, the packet of data which gets returned from that call is extremely verbose as it contains all of the information about each post.

Ideally, I would have options to retrieve either: A) just the id's of the post or B) just the post "headers". This way I could bring back some data that I could start working with very quickly and it would make it easier to attempt things such as re-tries on much smaller packets of data. For example, let's say that I do a GetLatestPosts(1000000) on Scoble's blog. I'm going to get back one almighty packet of data. Now, if I'm on a flaky connection and the connection drops out after I've *almost* retrieved all of the data, then I'm going to have to re-try again from the beginning. If I was able to request just the headers - or even better, just the id's - of his posts, then I would have that data very quickly and could start doing requests on just individual post entries.

The Atom API

After briefly chatting with Chris Frazier - who has actually had to implement some of this stuff - it was recommended that I take a look at the Atom spec. Having looked at the API documentation [2] it would appear that this spec. is obviously either incomplete or just in a very formative stage so I'll see what I can dig up on it over the next little while.

A better blogging API

I'll also do some searching over the next week or so to see what others have been discussing around this topic as it is obviously something which must have come up a lot. Maybe we have to wait until industry heavyweights start to build blogging products into their core platforms before clear spec's start to emerge around what is needed from an API that would allow for the programmatic disovery and interaction with blogs.

And then, who knows how long it will take for the current blog engines to implement that?

Links:
[1] MetaWeblogAPI: http://www.xmlrpc.com/metaWeblogApi
[2] Atom API draft spec: http://bitworking.org/rfc/draft-gregorio-07.html
[3] Atom high level: http://www.xml.com/pub/a/2003/10/15/dive.html

Another SUB installation

MarkItUp — Sunday, July 24, 2005 8:35 AM

Bruce Mack is a guy that I "know" through virtual collaboration over the past year or so. We "met" after Bruce got interested in ProjectDistributor and then wrote a smart client which uses the PD web services to expose all of the functionality.

Throughout that time Bruce has also spent time showing me the benefits of software patterns. Anyway, Bruce has now installed SingleUserBlog and is blogging. You can find his blog here:

http://blog.brucejmack.biz/

SingleUserBlog source code now online

MarkItUp — Thursday, July 21, 2005 7:35 AM

I've just released the first public version of SingleUserBlog which you can grab from here:

http://projectdistributor.net/Releases/Release.aspx?releaseId=220

SingleUserBlog is the blogging application which runs my blog (this blog) and currently supports a FileSystem data provider as it's data store. The main reason for writing this application was that I wanted something which was insanely simple - and I think that I have pretty much achieved that. If you download it and run it for yourself, please let me know so that I can keep tabs on it.

The plans are now to release a .NET V2 version of it on an account that I just set up with http://Orcsweb.com. I should have that online next week sometime I hope. The plan will be to release it so that others can go online to try out the admin features and see it all online and play around with it.

Added CommentAPI extensions to my Rss feed

MarkItUp — Thursday, July 07, 2005 7:02 AM

The Rss feed for SingleUserBlog now includes several extra nodes which make it play better with good Rss subscription tools. The new nodes include the wfw:comment and the wfw:commentRss nodes which provide links back to the location of comments on this site. Specifically the wfw:commentRss node allows your aggregator to do a lookup and bring the comments back to you so that you don't have to go to my site to read them.

I also added the slash:comments extension which contains number of comments made to the specified blog item. This is the thing which allows your aggregator to display the number of comments against each item.

You can read about the CommentAPI here: http://wellformedweb.org/story/9 and here http://wellformedweb.org/news/wfw_namespace_elements/.