Rename snippet library sub-directories

[tkphd]

At present, each of the sub-directories of each site's snippet catalog (to extend the library metaphor) is named with the numerical value of the lesson it goes with. However, as we see in #225, when new material is inserted, these numbers go out of sync. This is nightmarish to maintain, especially given the number of forks and pull requests that would need updating any time the lesson order changed.

Recommend finding a content-based naming scheme, with some verbosity, to improve maintainability & legibility. Related to #9.

Current

File Name	Title	Snippet Folder Name
11-hpc-intro.md	Why use a Cluster?	11
12-cluster.md	Working on a remote HPC system	12
13-scheduler.md	Scheduling jobs	13
14-modules.md	Accessing software	14
15-transferring-files.md	Transferring files	15
16-resources.md	Using resources effectively	16
17-responsiblity.md	Using shared resources responsibly	17

PR #225

File Name	Title	Snippet Folder Name
11-hpc-intro.md	Why use a Cluster?	11
12-cluster.md	Working on a remote HPC system	12
13-scheduler.md	Scheduling jobs	13
14-modules.md	Accessing software	14
15-transferring-files.md	Transferring files	15
16-parallel.md	Running a parallel job	pll
17-resources.md	Using resources effectively	16
18-responsiblity.md	Using shared resources responsibly	17

Proposed

File Name	Title	Snippet Folder Name
10-hpc-intro.md	Why use a Cluster?	intro
15-cluster.md	Working on a remote HPC system	cluster
20-scheduler.md	Scheduling jobs	scheduler
25-modules.md	Accessing software	modules
30-transfers.md	Transferring files	transfers
35-parallel.md	Running a parallel job	parallel
40-resources.md	Using resources effectively	resources
45-responsiblity.md	Using shared resources responsibly	responsibility

We would need to modify the lesson-check script to let increasing
numbering sequences to pass, otherwise tests will break.

[Sabryr]

It will be convenient if the snippet and lesson have the same name. For example:

File Name	Title	Snippet Folder Name
10-hpc-intro.md	Why use a Cluster?	intro

I would say

File Name	Title	Snippet Folder Name
10-hpc-intro.md	Why use a Cluster?	10-hpc-intro

Edited for table formatting

[reid-a]

@Sabryr I'm afraid that including the number defeats part of the purpose of the change. I'm on-board with:

File Name	Title	Snippet Folder Name
10-hpc-intro.md	Why use a Cluster?	hpc-intro

i.e. the snippet directory name could exactly match the text part of the base of the lesson file name.

@tkphd What are the modifications required to the lesson-check script?

Edited for table formatting

[tkphd]

Thanks for the feedback, @Sabryr and @reida. I tend to prefer @reida's suggestion for the first lesson, producing

File Name	Title	Snippet Folder Name
10-hpc-intro.md	Why use a Cluster?	hpc-intro
15-cluster.md	Working on a remote HPC system	cluster
20-scheduler.md	Scheduling jobs	scheduler
25-modules.md	Accessing software	modules
30-transfers.md	Transferring files	transfers
35-parallel.md	Running a parallel job	parallel
40-resources.md	Using resources effectively	resources
45-responsiblity.md	Using shared resources responsibly	responsibility

I think the easiest "fix" for lesson-check is to drop lines 299 to 307. This would retain checks for duplicate lesson numbers, but would not check whether the numbering increments by 1.

[tkphd]

Issue opened upstream.

[jevbelikov]

+1 for content-based naming; perhaps we could drop the number from file names too? this would also make reordering easier.

[tkphd]

Good prompt, @jevbelikov. If we want to, we can migrate to manual episode order, and name the Markdown files anything we like. This introduces a small maintenance & training overhead, since the existence and significance of the new episode_order array in _config.yml must be clearly explained. This would improve flexibility, though I'm sure there will be vigorous discussion of a new naming scheme were we to go this route.

[reid-a]

For what it's worth, notwithstanding the limited ambition of my PR, I am not opposed to nonsequential numbering, or to removing the numbers and doing manual sequencing. I left that out mostly based on uncertainty about how to coordinate with upstream on those issues.

[reid-a]

Closed in favor of #230