Example for Buckets
I am working on a database system called Buckets. The idea behind this are structured data records that are bound together through (optionally) classified associations. The goal of this is to have a structured and interlinked set of data records to organize information. Here’s an example of something I’m doing where I could use this system… I’m dealing with an external company to integrate their product with ours. Despite documented standards, there are always issues with integration. So I’ve started a Word document to keep track of issues that come up and how they are being resolved. Word isn’t the right tool. Excel isn’t either. And Access is a huge overkill as it will take more time to setup the database than will gain from having it. What I need is something in-between. Something that will allow me to capture pieces of data and have that data linked together in a meaningful way. Let’s look at what I’m doing now. The top of the document has a title and a short description of the purpose of this document. Really, this is an IntegrationDocumention record with two fields, Title and Description. Next I have a heading for each issue. I manually timestamp when this issue happened and have a description of what the issue is. In our database, this might be an Issue record with two fields, Title and Description. For each issue, there is an ongoing discussion of details and resolution options. Some of this is cut-and-paste from e-mails, some of it is snippets of documentation, some might just be off-the-cuff thoughts of what we can try or what we have talked about. There might be phone conversation notes (or audio) too. In this Word document, it’s just free-form text that sometimes uses sub-headings to help separate thoughts. In our database, all of these things might be stored in a variety of record types such as Comment, Note, Audio, Picture and so on. The problem with Word is I can collect all of this information, but there is no structure. There is nothing that separates one thought from another besides formatting. If you used Excel, you could create an issue spreadsheet with rows for each issue and columns for information collected, but it doesn’t allow for large, possibly interlinked, data to be put in there. A hack is the use columns titled (comment1, comment2, comment3), but that’s just lame and doesn’t really help. So going back to Buckets, we have defined the record types that we want to keep along with data in each record that we want to collect. The parent-child relationship is simple IntegrationDocumentation -> Issue. Issue -> Comment,Note,Audio,Picture. (where -> means has_many) We might want to have an additional level between IntegrationDocumentation and Issue for unresolved and resolved issues. Or an Issue might just have an association kind of “resolved” or “unresolved” that can be queried and displayed. In the former solution, you have a “bucket” of unresolved issues and a “bucket” of resolved issues. By virtue of an issue being in one of these buckets, it denotes both state and context.
