The Problem with PDF for Data

PDF is great, we love Adobe Acrobat. However, somewhere along the way someone got the impression that PDF is a good format for including tabular data. This is a problem because it locks up data. PDF is for creating documents that show the same on any computer and in any browser. They are not useful for and are in fact counterproductive for representing data. Also, it is unnecessary to actually embedd data into PDF documents as Adobe Professional has the abiliy to add attachments. This was you can bundle up both the text and the data all in one file.

What to Attach?

Excel is considered the standard file type. CSV is smaller and brings less baggage, but then only the raw data persists. However generally, the issue with attaching a file is it becomes out of date. If a link is included, if new data comes to light it can be added to the hosted file. 

Google Docs or DabbleDB?

If the spreadsheet is not too complex, you can add a link to Google Spreadsheets. However, a link to a DabbleDB form or view is also an option. A final option is to add a link to an RSS feed, which can also be pushed out by DabbleDB.

Multiple Field Categories

What is interesting is that DabbleDB has the concept of a single field belonging to multiple categories. This is different than a primary key relationship. We quote from DabbldDB help.

“default view for one category that also filters for the additional category or categories. Entries created from such a view will always be in both (or all) categories that are specified as filters.” - http://www.dabbledb.com/help/index/multiple_categories

Now this is the tricky part, createing a default view for one category that also filters for the aditional category or categories. 

On the help on making filters we find the following: 

“Multiple categories, you will have the option to add additional category filters. (The option will appear below the existing filters.) This will restrict the view to entries belonging only to both or all of the categories.”

We are stuck on this one. 

Schedule Update

Something very interesting is that any feed can be scheduled to be updated into the DabbleDB application. We start off by selecting the Import button.

Now we select the make schedule button and then select how often we want the feed updated.

RSS to GUI

Something quite interesting is that any dataset or category in DabbleDB can be setup as an RSS feed. This means it can be imported into blogs, web pages, etc.. However, What would be interesting is if there was a front end web based GUI that could be fed these RSS feeds. This would allow for a very rich application cto be created on the internet. Adobe Flex is designed to accept feeds or “web services” into a GUI. We are currently investigating Flex and will write more on the topic. However, our previous experience with Adobe products is that they have a difficult learning curve (a la Dreamweaver, although that was developed by MacroMedia), we don’t want to have the UI development consume all of our time and that is a concern with using Adobe Flex.

Background

DabbleDB used to support the Javascript API for connecting to a GUI. However, they changed their emphasis. See the exerpt below from the DabbleDB JavaScript API guide.

http://dabbledb.com/help/guides/jsapi/ 

“Important note: The Javascript API for submitting data to your database is currently disabled on most databases. (It remains enabled for reading data.) We no longer recommend using the API for submitting data. Instead, please try using the Pages feature, which allows you to build forms without any coding. Pages forms can also be used to submit data programattically. For additional information, visit the User’s forum or contact us.”

Export to RSS and Other

The export feature of DabbleDB really shows its commitment to the semantic web (a framework that allows data to be shared and reused across applications, enterprise and community boundaries – W3C) These links allow the same data to be viewed a number of different ways. The RSS link is an interesting one. Although RSS is used primarily for news feeds, there is nothing to stop RSS from being used to transmit large datasets. We have been investigating this site to understand how it supports RSS management http://www.rssbus.com)

You can see how several of these look by selecting the links below.

http://symmetricalcontent.dabbledb.com/publish/symmetricalcontent/bbe5613e-c97c-43c7-8be7-713e6c1ed3b0/calendarview.txt

feed://symmetricalcontent.dabbledb.com/publish/symmetricalcontent/bbe5613e-c97c-43c7-8be7-713e6c1ed3b0/calendarview.rss

Backup

The entire database in DabbleDB is saved as a series of CSV files (one per category). Creating, downloading and viewing the backup can be understood from viewing the screenshots below.

Filter

A filter is actually easy to create, although there is one extra character of syntax required above and beyond that which is immediately available.

To the right we need to include the “:” after the field name. Now when we select filter, it brings just those rows that meet the criteria.

Linking

Linking is one of the most important concepts in DabbleDB. A link between categories is similar if not identical to creating a link between two fields in an RDBMS table. Furthermore, a successful link creates the following entries as you can see below.

Here are the two categories that have been linked above.

Now we will see what happens when we add another invoice.

Scientific Databases

What is interesting is how difficult it is to find scientific datasets online. There is a tremendous amount of scientific research that is publicly supported every year. Publicly supported research falls into he public domain, and this raw data should be available for everyone. However, this does not seem to be the case. The raw data does not seem to be placed online. One exception is the site http://data.gbif.org/datasets/

There are many datasets as this sites that can be downloaded.

Integrated Data

Its really time for data to become integrated across the web. That is aside from RSS feeds, sites should have options to send datasets to other sites. DabbleDB has a facility to accept the data from other sites, however many sites do not have this data listed, even a site like http://data.gbif.org, so the other sites need to publish in a format that is readable to distributed databases like DabbleDB.

Example

Actually, you can see this. When we add the site CEPR.net, it brings in the RSS feed of recent articles.

Interesting DabbleDB Features

We have recently begun using Dabble DB. Some of the traits of DabbleDB are different from relational systems. This includes the following.

1. Categories are like tables
2. Unlike with traditional database tables, a single entry can belong to multiple categories at once.
3. A view is based upon a table or a form, chart or calendar
4. Columns can be used for grouping and sorting, as well as for things like subtotals and formula

This information was from

http://dabbledb.com/help/quickref/