Ensure the unique IDs assigned to each SUT in the results table are not changing when result spreadsheet is regenerated.

[arjunsuresh]

Currently each time the result table is regenerated the unique IDs of systems are changing if any new result is added or if some result is dropped. It will be good to ensure they remain consistent.

[arjunsuresh]

Proposal: Change of IDs in the Result Table

We are proposing to change the IDs in the result table as follows:

Current IDs

4.1-0001     AMD     .... (System1)
4.1-0002    AMD .... (System2)
4.1-0003    ASUSTek ... (System1)
....

Proposed Change (Step 1)

4.1-001-001     AMD     .... (Submitter1, System1)
4.1-001-002    AMD .... (Submitter1, System2)
4.1-002-001    ASUSTek ... (Submitter2, System1)
....

Explanation:

• We are introducing a separate ID component for every submitter. This component remains constant even when results are regenerated.
• The format allows for up to 999 submitters and 999 systems per submitter.
• This new format ensures that all submission IDs from a given submitter will always be contiguous when we preserve the already generated IDs.

Question: Should this ID be unique across different submission rounds?

Additionally, we need to maintain a record of previously generated IDs for each system (within the same round) so that when results are regenerated:

• IDs are first assigned based on the historical record.
• New systems will receive newly generated IDs.

Key Considerations:

1. Withdrawals: Corresponding IDs will remain unused.
2. New Results: Any new results (e.g., added by review committee decision) will receive new IDs that have not been used in previously generated results.

Advantages

1. Consistency: Immediately after submissions, submitters will receive IDs for their results that will not change based on withdrawals or additions by other submitters.

2. Reproducibility: Anyone can run the submission checker on the submissions repository and regenerate the results with the same IDs. All that is required is to save the previously generated IDs in the repository.

[mrmhodak]

Good suggestion. @mrmhodak to bring for discussion next week

[pgmpablo157321]

@arjunsuresh How should we handle when a system is renamed? I see a couple of options here, but not sure if they are a what we are looking for

• We can have a new ID when the system is renamed
• We can keep track the folder names, and assign the IDs based on that. That means, they wont change if the system is renamed, but when the folder changes. This might also happen during review
• Provide an option to map names when they are changed. Not sure how this will work, and if we have to do additional checks

[mrmhodak]

To be discussed 12/3