I recently hired into a data analytics team for a hospital, and we don’t have a style guide. Lots of frustration from folks working with legacy code…I thought putting together a style guide would help folks working with code they didn’t write, starting with requiring a header for SQL scripts first as low hanging fruit.
Or so I thought.
My counterpart over application development says that we shouldnt be documenting any metadata in-line, and he’d rather implement “docfx” if we want to improve code metadata and documentation. I’m terrified of half-implementing yet another application to further muddy the waters–i’m concerned it will become just one-more place to look while troubleshooting something.
Am I going crazy? I thought code headers were an industry standard, and in-line comments are regarded as practically necessary when working with a larger team…
You must log in or # to comment.
I recently hired into a data analytics team
I work in Data Engineering and have spent most my time on analytics teams. They don’t have a SWE/CS background and generally because of that don’t follow any good programming practices. In my experience style guides are hard to get them to follow properly even if you set up SQLFluff for them., I can barely make them see the advantage of not committing directly to main (at least we’re using git). It’s very frustrating.
Yep that’s us–maybe half of us have CS degrees.
The funny thing is that the pushback is coming from the “regular” development folks. At least we’re using git too :)
I think it depends on what you want to put in the header. An overall description of the file, it’s goal, etc - that sounds great.
A changelog saying who changed what, when? Our version control system is tracking that for us already.
Hmm, do I want to open some external site/program to see my documentation or have it already in the code in front of me?
We use doxygen at my company and I think I’ve only ever opened it twice in 9 years.
What do you mean by Code headers?
I hope you don’t mean those “Created by:” and “Last edited:” things… If yes, please don’t!
Commenting your code is an industry standard. Some kind of separate comment metadata won’t generally work, unless you code exclusively in spreadsheets, where you can add a note to each cell.
Given that it’s a hospital, I guess your management will listen to whomever performs bureaucracy better, not whomever codes better.
A header might be useful, although there’s likely better ways to (not) document what each sql statement does.
But inline documentation? I’d suggest trying to work around that. Here’s an explanation as to why: https://youtu.be/Bf7vDBBOBUA
If possible, and as much as possible, things should simply make enough sense to be self documenting. With only the high level concepts actually documented. Everything else is at risk to be outdated or worse, confuse
Self-documenting code only documents what the code does, not why it does it. I can look at a well written method that populates a list with random elements from another list and go “I know what that does!” but reading the code doesn’t tell me the reason this code was written or why alternatives weren’t chosen.
In the case of Rust, it goes even a step further when working with unsafe code. Sure I know what invariants need to be held for unsafe code to be sound, but not everyone does, and it isn’t always clear why a particular assumption made in an unsafe block (the list has at least 5 elements, for example) can be made soundly.
…what the code does, not why it does it
This is my issue with “it’s self documenting code!”. I’m a maintenance coder. I deal with people’s code long after they’re dead (or ragequit). Some are for control systems.
if (waterPressure_psi > 500) raise PipeMayBurstException. Okay, we’re dealing with water pressure, in psi unit, and if it’s too high, it may break the piping. Self documenting!!Except that our pipes are rated for 1000psi. SO WHY THE 500?! Do we have one or two sites - out of hundreds - with lower rated pipes? I can double performance if we raise the threshold to 700, well within the safety tolerance, but AM I GONNA KILL SOMEONE when they upgrade to our latest controller??
That’s because they are usuing magic numbers. If e.g. the 500 was MaximumPipeRating * SafetyMargin it would already be better.
If that list code is in a function called “PickRandomQuizQuestions” you would also know why it does that.
I encourage you to find a name for this function that describes why there is a second inner function. One restriction - the name of the function must be
run(that’s what the trait being implemented calls it, you can’t rename it).Sure, you can call the inner function
run_inner_to_fix_rustc_issue_probably_caused_by_multiple_fnmut_implsbut is that really any better than using two forward slashes to explain the context?
Your friend may have a point.
It depends where the SQL is.
Is the SQL in a data model in an analytics platform? Some platforms will happily carry comments around like last week’s pizza during query generative phases of visualization, so it may not be appropriate to put comments inside a data model, as those comments could become bugs if the analytics platform is lame, like most are.
Others, certain flavors of SQL DDL (Tables, views, etc), comments outside the DDL don’t make it inside the resulting object, so headers may not be the right place either. Most RDBMS have meta-descriptors that can apply to DDL so those might be good to look at.
For arbitrary SQL, outside a brief inline comment describing why it exists, and what invokes it, your next best bet may be a link to a more descriptive data architecture diagram that shows how this unit of SQL integrates with others. You might prefer hyperlinked descriptions from that data architecture over searching thru code.
As long as comments don’t require continual parsing (a one-time tax is inconsequential), definitely add details that you have figured out so others don’t have to re-learn the tribal mysteries of long-deceased ancients.
