improved lis queries for the weblog app

i added a handful of improved list queries for the weblog app today. these include the following

/data/weblogs/?recent - returns the most recent five entries

/data/weblogs/?category={@id} - returns all the entries for the requested category id

/data/weblogs/?year={@yyyy}&month={@mm} - returns all the entries for the requested year & month



i also added some list filters for the website list:

/data/websites/?recent - returns the ten most recent website entries

/data/websites/?category={@id} - returns all the entries for the requested category id



this will make building a typical blog home page easier and more consistent.



i updated the handler code and had to twiddle the isapi rewriter regexp, too. the rewriter update we more tedious then the code update!  anyway, after making changes to the code and rewriter, the /data/* calls all work as expected. my next task will be to define the 'public' URI patterns and implement handlers for them, too.



i also need to confirm that the framework code is using the MVP XML library - specifically support for x:include. i like the x:include pattern since it makes it very easy to build server-side pages that contain several 'xml parts.' this also means that i need to make sure other things like state and security will work with x:include. based on my past experience, this should go fine, but i need to dig into it to be sure.



anyway, things continue to move along well. assuming all goes fine, it's quite possible that i'll have the blog ready for data entry by the end of this coming weekend.

added support for the website objects

i added support for the website objects today. this will be my 'blog roll.' for the blog.  that rounds out my initial set of objects for the blog:

- weblog

- category

- website



i tested them all against the data api and all work as expected. i think i need to update the _list function a bit. maybe return a default top X number unless the 'all' argument is passed, i think. i'll add some other filtered list requests later - by year/month, by category should be good starters.



i think i need to add the top X return to the website_list, too. i'll proly work that out this week.



then it's on to the 'live' version of the site. for starters i'll define the following:

- home page

  + last 5 full posts

  + list of categories

  + list of websites

  + link to get full list of posts (reverse chrono order)



later i'll add some other stuff like:

- contact page

- posts by category

- list of all websites (full blogroll)



i need to implement a url pattern for the posts that work on a 'hackable' model:

- /posts/yyyy/mm/post-id - returns a single post

- /posts/yyyy/mm/ - returns list of posts for that month

- /posts/yyyy/ - returns a list of posts for that year

- /posts/ - returns all posts



this probably will be handled via the rewriter

/blogging/yyyy/mm/

/blogging/?year=yyyy&month=mm



then i just need to implement the filters for the _list query.



i also need to get a good editor working for the blog posts.

- edit weblog entries (uses a category list)

- edit category entries

- edit website entries (uses a category list)



well - that's all for now.

started my weblog implementation

finally got the courage to start to implement my weblog in the new REST framework <ta-da!>.



i was able to implement the database side (weblog, category, websites) in a little more than an hour. of course, i'd done some planning ahead of time so the implementation was just 1-2-3 simple. that included the table definitions as well as the standard sprocs that produce xml output from sql server (_list, _read, _add, _update, _delete).  later, i suspect i need to add some customized _list sprocs (most recent, search by date, etc.).



i was able to build each http handler in about 30 minutes. this involved copying an existing handler and modifying a few references to XSD and XSL files. then actually creating the XSD to validate the input and XSL to convert the XML input into acceptable T-SQL to submit to the database server. i got the weblog and category classes completed. i still need to implement the website class, tho.



also, i set up an endpoint for the weblog and the category classes. that meant updating my handlerfactory class and modifying the auth-url and auth-users xml files to control access to the new endpoints.  that took just a few minutes.



finally, i copied my stock html-ajax page that allows me to target the endpoint and do add/update/delete actions. this is my basic test bed. that took me less than 30 minutes to handle both the weblog and teh category. i was able to create sample data for both classes - no problem.



i also took the time to re-org the endpoints. now, all the actual data access is done through the /data/* URI. originally, i had the data behind each app (/running/runlogs, eating/foodlogs, etc.), but as the data library grows, i think it will be easier to control (esp. access control) if it all appears under a single /data/ endpoint (/data/foodlogs/, data/runlogs/, etc.). i'm not sure how deep to make the hierarchy - there's no limits, i guess. it will take time and some experimenting, but i'll get it.



i need to implement the website endpoint (handler and test page), but that should take less than 30 minutes. once that is done, i'll be ready to implement a 'live' display and edit service for the weblog.  i'm thinking that i will finally start to integrate the DOM/DHTML UI components that i scouted over the last few months (tabs, tables, tree, etc.).



anyway, it's been a good week. doing this next app has a bit more complexity and it's giving me a chance to expand the features and value of the rest framework. all fun stuff!

updated my food log app

i updated the foodlog app to look/feel similar to the runlog app i updated last week. it now has a slicker public side and a bit smarter editing experience. they bos support html content for the notes section and i'm starting to use things like anchors and image tags to spruce up the content - looking nice.



i want to run it a bit longer here on my workstation iis edition before packing it up for the public server. but i suspect the public server will happen before the end of the month. 



in the meantime, i am starting to plan out a simple blog app. the first pass will have entries and categories. i'm considering a blog roll and comments, but i might use an available service for those (get to test my ajax/rss chops!).  anyway, the app will be a bit more complex than the 'log' apps i'm using today. just the right step up in my process for building rest-ian apps.



on another front, i've stopped working on the http auth side of things. while basic-auth is working, i have no ssl cert to guard my transactions. i also haven't implemented digest auth. that is probably what i should focus on next. i also want to tighten up the session and cookie services baked into the framework. finally, the user accounts and url auth data is currently stored in xml files. that will need to be moved to a db relatively soon, too.



it's looking good and solid - i like it!

moved exyus runtime to root folder

made another big step today in rolling out a working version of exyus - moving it to the root folder of a server.



actually, a sub folder. but the point is that the app no longer runs under the asp.net debug server. i implemented some isap rewrite rules, modified the xsl and scripts and now all runs fine as a standard iis-hosted app. it's all about the configs now.



i tweaked my running log app to use a bit of javascript (running.js) and a clean css file (running.css). i need to do the same with the food log app. once that's done, both apps will be fully functional and stand-alone.



next, i need to package up the two apps and post them to my public-facing server that uses sql server (not the express flavor i have here).  at that point, the release will be complete - a fully functional rest-ian lightweight web framework.



once the thing is hosted publicly, i just need to start designing and implementing other rest-ian web apps. my next target - a blog app. then i can test my rss/atom output skills, too!

implemented a new 'app' in about two hours

had a good time this weekend. i implemented a second simple http/rest web app in about two hours. this time, instead of my jogging log, i created a simple weight loss log.  while it was really  simple (just one table with a handful of inputs) it was important to me to create a new app from scratch to test out the rest-server pattern io've been working on for the last month or so.



and it went well. i did do a lot of copy and paste (one reason it went so fast). that means i might be able to factor in some of the copy-paste into the base class. but i want to make sure not to 'over-engineer' things. i implemented the xml data uri, then a simple xhtml list/detail pattern for public viewing. setting up the security was a bit tricky at first (forgot some of my details for setting up secure urls and granting user permissions), but it fell into place once i dusted of the ideas. currently the security is all done via XML files, but i can see how doing this in a db would make things smoother.



finally, i also updated the running app by cleaning up the uri pattern and refreshing the public pages. it took tiny bit of work to update the XSLT and a bit of trundling the javascript client code. but it's all good.



i'm having some fun here. my next target? posting all this stuff to a public server instead of just on my workstation.

side track into client ui

i've been spending the last several days (at work) putting together some web 2.0 ui pages. this is focus on a POX (plain old xml) back end and ajax-inated javascript browser front end.  been kind fun, actually.



in the process, i've decided to build a small (but useful) DOM-friendly collection of ui components. i'm focusing on the following:

- tree control

- table (sort-able, filter-able, page-able)

- tabs

- menu (fly out and drop-down)

- modal dialog

- collapser (toggle page sections open/closed)

- auto-suggest (ala google, et al)



i also want to settle on a calendar and dhtml editor, but have not found anything i like yet.



also, i've worked up a basic 'behind-the-scenes' set of javascript libraries to make my life easier:

- dom helper (base from dean edwards)

- mozxpath (nice cross-browser xpath against xml docs)

- ajax library (built on the one from ajax patterns wiki)

- data library (found a decent one i need to work on)



so that's a nice set to starat with, eh?



so far, i've done enough work with the ajax, and mozxpath to know they work great. i like dean edward's base2, but it's a bit early to tell if this will last.



i also posted a clean modal dialog and auto-suggest set in my googlecode space.  i plan on posting the other parts as they get exercised (or exorcised, as the case may be<g>).



fwiw, i still have work to do on the server-side rest library. but that might wait a week or so as i get through this client stuff.