Project Update #4

[stanickld]

This week Sarah and I are focused on completing structural and descriptive markup of our song lyrics. We had a little bit of a slow start trying to figure out the schema, which is definitely going to need to be revised as we finish marking our lyrics up and making sure the elements we use are identical. I know for a fact the odd numbered songs have an incorrect element when it comes to the numbering system so that's something I plan on editing over the weekend. Relax NG is something both of us needed a refresher on and caused us to be a little bit hesitant to begin the steps necessary for markup. We also standardized song names in the style "songname-#.xml".

In our last project meeting with our mentor, we discussed giving someone the job of physically drawing what we wanted our website to look like; hopefully by next week I will be able to confer with my partner about what we want the layout of the sight to look like and then complete sketches from there. We're both a little bit concerned about the time table for our midterm goals as we got a late start to completing our markup. Sarah is completing markup on the even numbered songs and I have the odd numbered songs in order to split the work up evenly; the hope is that even if we need to do revisions on our descriptive markup, we'll at least have a solid foundation for what we've completed so far.

[ajm324]

My group also took the approach of editing the schema as we go through the mark-up process, which is especially helpful when working with a lot of text because it is hard to know every device you will want to mark. We focused less on standardization at the front-end, even though we all followed the schema, and will have one person review the markup for consistency and accuracy. This is super helpful for us because the text is in a different language and I have the most experience to check the mark-up, but maybe a similar review process would be useful for you if consistency is a worry. It's also good to hear you guys are starting to think about website design, the physical drawing of layout seems really helpful and I may have to try that over the weekend!

[gec34]

Our group’s schema has also largely been subject to change. Each group member added new elements and attributes as the process has continued. This was helpful working with a large amount of material to markup just in case there were possible elements more prevalent in one section. Having a super strict schema before going into markup might mean you miss a possible pattern so I think you are making good progress on it. Having a standardized song naming system is a good idea to help keep your directory organized.

[djbpitt]

Ideally we perform perfect document analysis before we type a single angle bracket, identifying everything we care about in our documents. Then we write a perfect schema before we type a single angle bracket, ensuring that our markup will allow only what we want it to allow, and nothing else. Only then do we begin tagging.

In practice it almost never works that way because our initial document analysis is based on a sampling of our documents, so we won't necessarily have seen or noticed or understood everything we'll need to mark up. For that reason, the process above is more of a loop, where discovering new things during tagging is essentially the next iteration of document analysis, which leads to a revision of the schema, which leads to a revision (including retroactive revision of documents already tagged) of the markup, etc.

Although the practice is almost always iterative, it's still a good idea to foreground document analysis at every stage. Sometimes revising a schema can require a lot of adjustments, and then a lot of adjustments to documents we've already marked up, so the better the document analysis (at any iteration of the loop), the easier the development.

[clk125]

My group has also been spending time updating our schema. I like your system for standardizing song names and I think that that could be helpful for my project as well. We were convinced that our super strict schema we originally came up with was good but as we each began marking up we realized that there were plenty of changes we would need to make.

[EdwinRosenberg]

This sounds almost exactly like what our group is doing. We are also finalizing our schema and making sure our texts are valid against them. It's always a bit of a fluid process, because you don't want your schema to be too strict in case a group member wants to add in a new element. On the other hand, you want all members to be suing the same elements, so its a balancing act. Standardizing song names seems like a great idea, we did something similar in our group as well. And good luck with the website! Our group hasn't even really talked about that yet.