Month: June 2003

this is a test post from mozblog

Mozblog allows me to connect directly with blojsom using the blogger API over xml-rpc. Its kinda of works but doesnt. As far as I got on my screen remotely I cant read my posts because it ignores all html files and only sees text files. So I could change my html files to text i guess.
Also when I hit publish, it throws a error, but still publishes it. Very weird error about objects…
Mozblog does support per entry metadata, so I can now start adding more details to each blog, expect more in this area soon…
Oh and it supports not only BloggerAPI but Metafilter and a couple others. So you can use it for anything from Moveabletype to Blojsom.

A few moans…
It does allow you to ftp your images to the server if it does ftp, but why not also support webdav?
And what makes it difficult for me to even think about deploying this in ravensbourne is the server address, username and password are all tucked away in the settings. So unless we can get moz reading its preferences from LDAP, basicly it wont work in the college.

Ah ha, I've been playing with microsoft's infopath beta.

So whats the crack?
Well its a form based maker for filling in xml documents. It uses a datasource as the start of the document and you can then build forms to fill in that datasource.

In english – it takes a schema, can be database or xml schema and creates forms for each element in the schema. Sounds good eh? Yes and it actually isnt that bad, it does what it says and within 1 hour I had a form which would fill in a courselet for me.

First problem
it hates my schemas to death. If you import or include a schema in or apply min or max to a choice it goes nuts. Which means you end up writing a schema just for infopath or simplifying your schema to the point it becomes non useable.
My orginal schema for courselets was fine, but the one for courselet 2 which included simple xlink with its namespace and a generic xhtml schema. Threw a error each time, i even tried to fix the errors and ended up making all my current xml documents un valid.

Next problem
Even if like me you make a custom schema just for infopath, and create the forms.

You have to own a copy of Infopath to fill it in! I thought you could open it in word or even ie6 and fill in the forms. No it seems like you must own infopath to save the final xml document. That sucks big time.

Plus points
It allows you to write validation in javascript or vbscript. It also understands xml schema and points out required and optional elements. I havent really tested the attributes too much because all my new schemas dont work in infopath. It also reckonmends how you should layout certain forms based on there schema type. So it will not create a freeflowing block of text for intergers. Or a drop down box for free flowing text.

It deals with repeatable elements better than I actually thought, everytime you press return it will create a new paragraph in my case. You can also change that to shift return or anything else you want.
Its also creates accessable forms using access key and tabs points, which is useful for the future. i have yet to try the xsd:any element and serious xhtml style schemas which allows the author to play with the format. You could allow them to put in element of there own using the greater than less than thing but you shouldnt have to. I would also like the ability to split forms up, so one form could be for dublin core metadata and the other for the xml content. It would then join them together using xinclude or something like that.
because even a form for a courselet looks big and most of the metadata is already filled in. yes you can do prefilled and subs. It will take defualts into consideration too.

So I would say its good but great for simple structured documents where the author doesnt have the freedom do what they like. For creative input, basicly forget it, its far too restrictive. But lets not forget its a beta and things will hopefully change.

Ok I have to confess Streamliner on the ipaq is actually quite good for writing down notes quickly. So good I might actually buy a copy.
One thing it doesnt do at this moment is, save or open opml files.
It does however save as wiki format and rtf. The rtf unfortually is styled and not in bullet lists as I orginally thought it would be in. The wiki is xml based and is actually valid, so I can always write a xsl which will take the node element and nodetext and turn them into a outline with nodetext becoming a attribute.
For some unknown reason it also puts the DTD for the wiki at the top of the file. Useful? unsure yet.

I tried to put the wiki file from streamliner into Java outline editor and it refuses to read it correctly. Instead it just bungs the whole lot into a new outline. So you can see all the node tags.
After futher experination I realised that Joe is expecting outlines not nodes, even on wiki files. Which is a shame because I'm gonna have to write my transformer soon if I want to keep using them.

I'm hoping the next version of streamliner will support opml natively or at least import/export. Or I can get joe to read nodes like outlines. That would be hardwordk.
I need to find out what officially is a wiki file, I guess streamliner has it right and joe is just being silly.

Its quite interesting that outline is actually a really good way to write structed notes quickly and easily. I never used streamliner before I went to the semantic web talk the other day. But I was able to write quick notes on my ipaq very quickly. And they actually make sense afterwards.

I really can not see one note beating this for speed and structure. Plus the fact theres editors on every platform.
I have yet to play with one note because it crashes everytime I run it, has to be something due to the fact my tablet feature doesnt work no longer. Will have to try it out on a non tablet machine see if it still crashes.
But going on what others have said, one note stores notes in its own format and it is not xml based in anyway. I guess Microsoft were thinking that you will bring your notes into word or something else and export as xml then.