User Story: RT-DC

[B-Hartmann]

The software should be easy to use and taking care of all the necessary post processing steps:

• compression
• renaming data
• copying data

It should do as many tasks as possible with as little user effort as possible, otherwise people will not use it.

First GUI draft

I imagine a user interface where I can see an "Input" area and an "Output" area. In the input area, I can just drag and drop a folder with all my measurements. The software then detects automatically by the file suffixes that it's about RT-DC data and selects the RT-DC folder on the network drive as target directory. This default output directory can be changed in the settings.
The software should copy all files in the chosen input directory (no suffix excluded).

There should be options to compress and copy the data:

• "as is": It preserves the folder structure and file names of the original data, it just compresses the data and copies it. This should probably be default.
• "change file names only": It preserves the folder structure but renames the files according to a user defined name plus an incremental number like "_M001", "_M002", etc for .rtdc files
• "change file names and copy in same folder": It will rename all files with a new unique name according to a user defined string and output everything into one folder. This might be done completely automatic and the unique names could include the following information (as mentioned here):
  • at least measurement number (in the style M001 - M999) for .rtdc files
  • date
  • flow rate or reservoir tag
  • sample name (derived from folder name)

Maybe useful for advanced users but too much for a standard user would be to make variable substitution usable, for example, with a string like %date%_%folder%_ and the option to copy everything into one folder, I would expect something like this to happen:

The new file names would start with the date, then comes the folder name of the folder they were originally saved in (usually this contains important information about the sample) and then there would be an incrementing number.

I can see in the output area already a preview of the processed data and it's folder structure to check if the output has the desired format.
I just click on a button "Start" or "Copy data" and the data is being processed.

Important questions

1. What happens in case the program gets closed (crash, user, blackout, task stopped via task manager) during a file transfer?
2. What if the destination folder gets disconnected during copying?
3. What if the software is closed during copying or maybe during compression, then the destination folder gets deleted and then the software is started again?
4. When renaming data: Should the sample name in the RT-DC file(s) also be updated or not
5. Should/Could DCOR-Aid functionality already be included?

[B-Hartmann]

A very cool feature would be if I could add extensions/plugin features so that the newly generated .rtdc files would already contain those new features, similar to how you can add them in Shape-Out2.
Pro: It could save me waiting time later during the data analysis.
Con: It would require more time for immediate processing and copying of the data.

[paulmueller]

Thanks for the detailed story!

These are the tasks into which this can be broken down

• Handle copying/disconnecting/crashing properly #2 (This belongs to the core, so I will work on it)
• Add initial GUI for RT-DC data #3
• Add logic for renaming RT-DC data files #4
• Add option to compute ancillary or plugin features for RT-DC data #5

I also have a general question:

Would it be better if the software "watches" a directory or if the user drag-and-drops files into the software?
There is a fundamental difference that has implications for automation.

Scenario 1:
Like in DCKit, users drag their files and MPL-Data-Cast finds the data location on the server (based on a default location for RT-DC datasets). The user might miss files here.

Scenario 2:
The user is presented with a directory tree where she can mark different folders or files for transfer. The user might be overwhelmed by the available files (but you have that in Windows explorer as well).

[B-Hartmann]

Scenario 2 - The software should watch a directory.

[B-Hartmann]

Regarding recipes and RTDC data: As default, the software should copy all files, not only .rtdc files. Maybe in the settings, one can select one of the available different recipes or add and select custom ones. Or maybe even chain them? But I guess that's not needed, since you can write your own recipes.

[B-Hartmann]

Scenario 2 - The software should watch a directory.

Actually, I was having second thoughts and think Scenario 1 would be better. I have the feeling this would make it easier for the user and that the chance of missing files is quite low.