Questions about Arctos as a community and collection management system
Q: What do you mean by the “Arctos Community”?
A: Arctos users form a community whose input directs development. Significant changes to Arctos require discussion among users. Such discussion, and posting of new issues, are communicated via the Arctos Google Group and the Arctos Working Group. Issues are posted, prioritized, and tracked on Github.
Questions about Arctos System Details
Q: What’s a VPD?
A: A Virtual Private Database (VPD) allows us to share resources, like programmers and hardware, along with some data, such as Taxonomy and Agents, while other data are masked from view or editing per data user or role. We all end up with more than we could afford by ourselves, and operators generally can’t tell that they’re in a shared environment.
Q: Why Oracle and ColdFusion?
A: Because they work. We’ve tried many other solutions along the way. Oracle is rock-solid and stable, and allows us to do things like share/control data via VPDs, maintain current data to our query environments, and sleep at night. ColdFusion is a very robust rapid development environment that fits our programming style perfectly while providing very close to 100% uptime and reliability. On a more practical level, implementing an open-source solution would necessitate hiring at least one additional person to manage software, while compromising stability and security.
Q: What about security and backups?
A: Arctos has multiple levels of security through its collaboration with the Texas Advanced Computing Center (TACC). A lightweight application security package controls access to forms, while Oracle partitions data by user, roles, and context, and provides auditing. Incremental backup logs are maintained on mirrored disks, and daily backups are maintained in three geographically separate secure locations.
Q: Are there specific system or browser requirements?
A: We attempt to keep the client-side of Arctos applications as generic as possible, but we have made some exceptions:
- Pop-ups: Users may wish to enable pop-ups. Some informational windows use pop-ups. We promise to only “pop up” things you ask for. Operators must enable pop-ups. Many browsers block this, sometimes cryptically, by default.
All applications have been tested in Firefox, and we highly recommend that all users upgrade to the latest release of Firefox (available from Mozilla). Most of Arctos should work most of the time in most other browsers. Let us know if you have trouble accessing this site in your browser, and we’ll fix it if we can.
Questions about Arctos data and functionality
Q: Are these live data?
A: Almost. Arctos data are stored in a highly normalized relational structure, which is great for organization but not so hot for query. Some data are then optimized for query performance by way of Database Triggers. Presentation data are generally less than 24 hours stale.
Q: Is there a limit on the number of records I can return in a search?
A: We impose no strict limits. Queries almost always take less than 5 seconds, although getting the data to your browser may become a bottleneck. If you have a reasonably fast browser and connection, it should be possible to return at least 100,000 basic records with a single query. Let us know if you find something excessively slow.
Q: I hear Arctos is complicated. Is that true?
A: Arctos is complex, as are the data it strives to accurately represent. There is a steep learning curve to understanding all functionality. Basic functionality – such as that available from other collections management systems – is pretty simple, and we think we do a pretty good job of making it intuitive. Perhaps more noticeable is the level of precision required to use Arctos. Rather than (mis!)typing a string, you may have to pick a value from a list, or you may have to supply metadata qualifying your assertions. We strongly believe that this is a necessary part of managing the specimens and data with which we have been entrusted, and thus worth the effort to learn this system.
Q: Does Arctos store GenBank sequences?
A: Arctos does not store GenBank sequences, but it does link out to GenBank if the GenBank number is entered as an other identifier in the specimen record. Conversely, GenBank will link back dynamically to the Arctos specimen record if the three-part GUID (institution:collectioncode:catalognumber, e.g., MSB:Mamm:148506) is entered in the GenBank specimen_voucher field. Here is an example of an Arctos specimen record with GenBank links.
Q: How is taxonomy managed in a shared system?
A: Identifications (optionally) reference taxonomy, but taxon names are not directly applied to specimens as Identifications. Taxonomy, therefore, strives to be “pure” names, leaving the complexities of applying names to specimens in Identifications. This arrangement:
- Removes ambiguity in determining what is “proper” taxonomy
- Allows more formality than other systems
- Allows us to use external sources of taxonomy
Any namestring (“taxon name” or “scientific name”) may have any number of classifications (“higher taxonomy”), and these may be drawn from Globalnames.org. Arctos taxonomy is therefore a synthesis of all “taxonomy projects” that contribute to GlobalNames, plus local data. Curating taxonomy in Arctos is theoretically unnecessary, although as of this writing all collections continue to do so due to the lack of responsive and useful editing capabilities in the various remote sources.
Q: Can I store images or video in Arctos?
A: Arctos defines media loosely as anything that can be linked through a URI. Internet-accessible images, sound files, video, web pages, and documents are all potentially media. Media may form relationships with any “node” in Arctos.
Arctos media are stored along with other data on servers administered by the Texas Advanced Computing Center (TACC). This partnership provides us with essentially unlimited storage space. Users with media-based projects need to create a TACC account that allows them to upload media through the TACC portal (see How to Upload Media to TACC). Arctos currently links to around 10 Terabytes of media.