week2+

Table of Contents Second Week for unicorn project : Formatting the tasklist To deal with the observer that will check the mediatypes and doctypes Mixing up all the ideas to create the bright one Creation of the logo

Second Week for unicorn project : Formatting the tasklist

Possible changes to make to the tasklist.xml in order to enhance the flexibility with conditionnal checking.BR We need to be able to have simple algorithms to define priorities in the tasklist.BR These nodes will be placed into the task node for each taskBR Here are our first ideas about the new nodes to add. BR The observer parts will remain unchanged for now except for the priority attribute.BR We can delete the priority attribute in handle mimetypeBR

A routine node to define the sequence to execute for this task :BR

• Conditions are called using XPATH and refers to condition section
• exec will execute the observer specified. As we call it with XPATH we can define a observation or a task -> inclusion of tasks -> recursivity

This routine node will replace the current subtask node.BR We could have several routines : for example for the general conformance which would check :

• semantic web
• page weight
• browser compatibility
• orthograph checker

We could instance some priorities on these check in order to make different subroutines depending on the preference An algorithm that first cares about page weight for the busy server. Another with semantic web... and so on! Dunno if my example is still good enough but i'm trying hard here!! :P

Another idea proposed by Jean-Gui would be to create a switch keyword to deal with doctypes and mediatypes. This way we could create different routine depending on the type we find. Therefore the condition would be in the if tag :

The if tag might become so big when we want several doctypes and mediatypes. However the switch might be an option to consider seriously :PBR You can see the third idea to have more details. BRBR A cond will be placed here to be used in if command. It returns a boolean :BR

• id to help calling the condition
• result to know if we want it to pass or to fail
• the list of doctypes that we want in it
• the list of mediatypes we're looking for

Conds can be used in order to do the multi-threading. In the previous version there was 3 types of priority. To insert the multi-threading they just ran all the tasks which belonged to the same priority in a map. As the maps will disappear in the new version as well as the priorities, we thought this node conds would be an good way to deal with multi-threading. We can multi-thread easily all the sub-nodes defined in conds. BRBR An outputseq will be placed to know the order of the output result display sequence :BR

• id to give the order
• and the name of the observation to display

This is to improve in order to check that it's well formed. For instance we shouldn't ask to display an observer that haven't been called due to conditional checking.

---

To deal with the observer that will check the mediatypes and doctypes

We thought we could deal with it as it was already done using a new attribute for a taskBR

---

 = Second idea =

We could re use the current priority properties. Let me explain: BR We 'll consider the if condition as a high priority and what's in the condition , lower priority. BR We can then build a whole map of different priorities and execute it as it's done currently.BR Of course we 'll need to extend that idea to take into account the mediatypes and so on.BRBR As we discussed it appears that this solution won't be possible. We can't keep the independance of the task and the mutiple 'if' insert .

 = Third idea =

This could be factorized to use task inclusion:

And in an HTML task:

BRBR

---

Mixing up all the ideas to create the bright one

BR

BR BR See the attached files for an example of the new tasklist BR BR

Creation of the logo

Jonathan created a logo in his free time , check the attached files ;)