|
It was only useless with the advent of more sophisticated compilers.
I started coding in 1986 - old habits.
And for the record, I *did* work for microsoft.
But more, when you're writing company code that's fine.
When you're writing APIs to expose outside of your company you should probably adopt standards for your public interfaces that are widely understood.
Anyone writing in .NET is going to be familiar with the style of the 6500+ base classes used in the .NET framework, so it's as good a standard as any for public facing APIs, and better than most
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
|
That seems like a childish response, tbh
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Seriously though, when you get older you live and let live. Truths become less absolute.
|
|
|
|
|
I agree with that. In most aspects of my life I am very buddhist.
In most of my coding I am. Public facing APIs are a hill I'll die on though.
And by public facing, I meant interfaces that are intended to be consumed by a wider audience than one's company/team
Standards here are sooo important, for reasons I could practically write a book on.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Well, do you think any API became unusable because of the style of its interface?
|
|
|
|
|
I think what happens is people get less inclined to use it as the learning curve grows.
Especially with the rate people are expected to learn technologies.
The more one-off your style is, the more difficult to adapt to it.
Furthermore, when integrating many components it makes the glue code clunky if the APIs don't even remotely match up to each other in terms of how they're named and how you navigate them.
Pretty soon everything is spaghetti. Glue has always been glue code, and somewhat messy, but how messy largely depends on how standardized everything is in terms of how it's programmed against.
It's hard to put a point on specifically, if only because so many APIs with a one-off style suffer from other problems (libutp is a good example) but problematic APIs lead to less adopters.
So if you want wider adoption and higher quality code it's a good idea to keep your public APIs playing nice and consistently with as much else as you can.
This means in .NET for example, at least familiarizing oneself with microsoft style and standard guidelines.
Why? because the 6500+ base classes in .NET are coded out that way, and people already know how to use them. So make your API similar and the learning curve is less steep - seems obvious to me but then i've been coding for a long time too - not quite 40 years like i guess some people here
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
I am an old programmer. started in 1978.
ASM all the way till 2010. yep old school.
In 2016 I picked up a gig with a company I left in 1992.
Tracking equipment.
All C based.
No problem.
NXP processor. Not bad.
MQX RTOS. Bloody night mare. (not an RTOS by the way)
Had to drill down and re-write the drivers and part of the OS.
The young buck in charge! did not like my style of programming.
I did not like his. (macro use every where)
Who cares if your style is not liked by someone else.
does the end result work? Can someone else follow the code you have produced?
If so then job done.
Am fed up with programmers trying to copy someone else's style and spending months making it look pretty instead of making a product work using their own style. (if possible)
STOP copying and pasting. Without understanding that which you are copying.
Anyway. The message is. Develop your own style and progress with that.
Ken.
|
|
|
|
|
That seems sensible, with a heavy caveat.
Public APIs with a general/public development audience (meaning beyond your company/team)
Like if you're making libraries or components.
At that point, I think it's important to adopt as widely available/common denominator standard as possible for the public facing APIs at least.
This makes everyone's lives easier, except maybe yours (or mine) due to the research and learning curve, but in the end it saves user headache and in some cases can cut some of the tech documentation down a bit because you don't have to re-explain basic concepts, like say, a dictionary object. If you're using those to enumerate keyed items you don't need to document an "IndexedCollection" class or whatever.
So in general i agree, but again public facing APIs are my sticking point. I'll die on that hill.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
The point I was trying to make is.
Do not loose heart if someone does not like your style.
Everyone has their own style.
So long as others can follow your code, does it matter.
As a programmer (if you view it as I) you will try to make it as efficient as possible. (think and re-write, think and re-write)
Keep a problem area in the back of your mind, it may be working, but you know there is a possible problem there.
Do not! copy someone else's style. (this is for the new comers)
Make mistakes.
Admit it and fix them.
A programming style cannot be learnt from a book.
I love using multi dimensional arrays. They are efficient, but the code is really hard to follow.
Therefore, I have to write a document explaining exactly how the code works.
Rambling, I know.
Ken.
|
|
|
|
|
> I love using multi dimensional arrays. They are efficient, but the code is really hard to follow.
In .NET they're inefficient. Nested arrays are more efficient. Weird, I know.
But in .NET I end up creating really ugly structures for a similar reason as you using MD arrays (like in C or C++)
Here's a recent datatype I had to munge through. Yes, I use comments for this
Tuple<TAccept, Tuple<KeyValuePair<char, char>[], int>[]>[]
Fortunately, it's opaque to the end consumer. It's basically a "handle"
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
I do would not like to come across that in a piece of code.
|
|
|
|
|
me either.
but it was either that or require a dependency in the generated output that i didn't need (it's part of a code generator)
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
I used to have OCD some years ago and I overcame it by noticing that I have it, noticing what particular actions I'm doing due to OCD and realizing that those actions are nonsense that cost time, don't even bring enjoyment and that's about it.
The rest was a tad of willpower and utilitarism.
|
|
|
|
|
Intellectually I understand that. In most areas of my life I can even apply it.
Here I'm battling 30 years of NOT doing that (being mindful and letting go) where it comes to coding.
Maybe I'm lacking in the willpower department. I usually find ways to hack around my lack of willpower. 
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
How much of a hardliner utilitarist are you? I very much am and that helps a huge lot in overcoming nonsensical habits.
I wouldn't even say that I'm big in the willpower department, but my utilitarism is a suitable substitute.
|
|
|
|
|
well i am an instrumentalist "what's real is what's useful"
and by real, i mean what exists in any meaningful sense
i am pretty much a hardliner about it maybe, if you'd consider that hardline.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
If you are reviewing code, use your smart editor to format it the way you want so you can understand it.
If you need to make a change after review, revert the file, make your change, test, move on.
And follow this rule:
Rule #1 for source control sanity retention:
Never combine a reformat with an actual code change in the same commit!
Use a separate commit with a "//reformat" comment as the first line for formatting, then remove the comment and do the real change.
|
|
|
|
|
The only strong conviction I have related to coding styles is when changing someone else's code.
It is expressed by:Quote: When in Rome, do as the Romans do.
It annoys the elephant out of me when a source file has 4 different styles
because people can't be bothered to conform to what's there or are
so stubborn they want to impose their style on everyone else, or they
simply don't give an elephant about anyone else.
/endRant
I need to stop reading these threads before coffee.
Cheers,
Mike Fidler
"I intend to live forever - so far, so good." Steven Wright
"I almost had a psychic girlfriend but she left me before we met." Also Steven Wright
"I'm addicted to placebos. I could quit, but it wouldn't matter." Steven Wright yet again.
|
|
|
|
|
codewitch honey crisis wrote: I need to be able to use other people's code without feeling a little sick about it, or wanting to refactor it before I touch it
There's an easy cure for this.. work on a project with a lot of devs that has a too-tight schedule and do as you want (i.e. shamelessly refactor all their code to your liking) until you wake up one day and realize that you're the one that needs to be duct taped to their chair so the product can ship on schedule.
Ok, I'm half serious about that answer. The real answer is that you need to realize you can't do everybody else's job for them, and accept that, while they won't do things the way you would, the things will get done well enough to not matter.
|
|
|
|
|
I have. I used to PM even. Code review is good, but if deadlines get squeezed a lot of stuff goes by the wayside, including reviews, and quality control in general
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Some companies put code through a pretty-printer as part of the check-in process. You could try that. But you have to be the first developer...
Boy...constants on the left...I haven't seen that in awhile.
I've been accused of having an "idiosyncratic" coding style because I like to add spaces to visually line up parts of my C++ class definitions. I don't worry about it. It's my practice, and I do what I want. If someone wants to change it, they can. If the coding standard actually says not to do it, then I don't.
Speaking of which, I once had a colleague edit my C++ code to change all the C++-style // comments into C-style /*...*/ comments. If you're reading this, "Whatever, dude."
|
|
|
|
|
My own experience was 1 year of working with very strict coding style guidelines, out of many years where things are much more laissez faire. Literally hundreds of formal rules, plus as many unstated informal one that everyone was supposed to know. Code reviews ended up being a bitch fest where 90% of the argy bargy was about code style. It literally slowed productivity and momentum to a crawl.
On all my other team projects, productivity was much higher, and not necessarily any worse technical debt-wise. Code reviews were simpler and much more productive/pleasant. There are some things one simply should not do (eg exception unsafe code, unnecessary pass-by-value) - everything else, just let it go man. I got quite tolerant of Hungarian notation, for example, just ignoring it as white noise (as it effectively is). You can tell instantly who wrote a piece of code from its style, so any questions, you know who to ask, if they're still around.
|
|
|
|
|
> You can tell instantly who wrote a piece of code from its style, so any questions, you know who to ask, if they're still around.
You make an excellent point. It's as good as a ID card sometimes.
For awhile I worked reversing worm code for a security outfit that was doing IT forensics, to find common authors of the code.
Even reversed from machine code, devs leave their mark.
I dropped hungarian notion
Right now in C++ I use a hybrid of different styles depending on what framework(s) I'm working with.
I don't like to do it, but the alternative is worse.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
She seemed surprised.
Already coated and running...
Software rusts. Simon Stephenson, ca 1994. So does this signature. me, 2012
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
