BF Workbench is a restricted scientific system. Records are available only to registered, authorised accounts.

Projects and datasets explained

What each is for, codes, status, and how records belong to them.

Reference

Every record on the platform lives in exactly one project, and may also be listed in any number of datasets. A project is where records are created and owned; a dataset is a named selection of records drawn from one or more projects, used to group them for a case, an analysis or a publication without moving them. Together, projects and datasets are called recordsets, and most of the workbench (the console, sharing, analyses and downloads) works the same way on either.

How it works

A project is a container with an owner. Records are uploaded into a project, edited within it, and can be moved from one project to another. A record cannot be removed from its project; it can only be moved to a different one.

A dataset holds references to records that already exist in projects. Adding a record to a dataset does not copy or move it: the record stays in its project and gains a second home. Removing it from the dataset drops that reference and leaves the record untouched. A record can be in as many datasets as you need.

Each recordset has its own member list and access levels, so a dataset is also the usual way to show a selection of records to someone who should not see the whole project. See Sharing a project or dataset.

Project Dataset
Holds The record itself A reference to a record
Per record Exactly one Any number
Code 3 to 7 letters, for example RKNZG 4 to 8 letters or digits, prefixed DS-, for example DS-UATDEMO
Owner level Project Manager Dataset Manager

Codes

The code is the recordset's permanent identifier. It appears in the HQ lists, in the top-bar jump box and in the console address (/recordset/<code>). Project codes are letters only and are stored in upper case. Dataset codes are letters or digits; the platform adds the DS- prefix itself, so enter only the part after it. The hyphen is how the workbench tells a dataset from a project.

Status

A recordset is normally private: only its members can open it, and only its members' records are returned by search. One status changes this. A recordset marked remote is one that was imported from another system. Every account on the platform can read its records through search, whether or not they are a member. The console of a remote recordset still opens only for members.

Your recordsets on HQ

The HQ page lists every project and dataset you belong to, in two tabs, with the number of records, records with sequences and sequences, and when it was last opened. Click a code to open its console. You can also type a code into the top-bar box on any page; it is upper-cased and opens if you have access.

Good to know: "Last Accessed" is the last time anyone opened the recordset, not the last time you did.

Create a project or dataset

Any signed-in user can create either. You become its Project Manager or Dataset Manager and are taken to the new, empty console, which shows a getting-started guide until the first records arrive.

  1. On HQ, use the new project button, or open Management in the sidebar and choose Create New Project (HQ only) or Create New Dataset (any page).
  2. Enter the code, a title and a description.
  3. Submit. The code must not already be in use.

Good to know: Type the code in upper case. A lower-case code is created correctly, but the page you are sent to afterwards reports an error; open the recordset from HQ instead. Avoid &, + and % in the title and description, as they are not stored correctly.

Delete a project or dataset

Only a platform administrator can delete a recordset, from the project and dataset lists in Platform Management, and only when it holds no records. There is no delete option on the console or on HQ. If you no longer need a dataset, remove its records first and ask an administrator; a project's records would have to be moved to another project before it could be deleted. See Moving records between projects and datasets.

Request publication review

The Publish Data menu on a console lets a manager submit a publication request for the recordset, with a title, authors, a statement of intent and two release acknowledgements. Any member can open Review for Publication to see the request's status, and the submitter or a manager can withdraw a queued request.

Limit: Approval of a request records a decision; it does not publish anything. The recordset's members and visibility stay exactly as they were.