This is a guest post for the ComputerWeekly Developer Community automation collection created in entire by Tim Vereecke in his purpose as principal internet overall performance architect at Akamai Systems – the corporation is recognised for its content material shipping network, cybersecurity and cloud support technologies.
Vereecke writes as follows…
It doesn’t issue if you’re a big team or a lone freelancer, automation is important in software program progress.
But that becoming said, there are quite a few all-way too-widespread pitfalls that progress groups tumble foul of when utilizing automation in just their workflow.
I’d like to discuss four of the most widespread pitfalls in automation-assisted software package growth and reminds viewers that the standard concepts of great software program advancement implement just as considerably (if not additional) in automation initiatives.
#1 Interface disgrace
[We often find that developers fall foul of ] losing time rebuilding current interfaces.
It commences with developers introducing a good new automation characteristic working with an API. But then the business builds a lot more and extra features on top rated of this, far more choices, extra attributes, far more flexibility.
Prior to you know it, you’ve nearly rebuilt the present solution or UI – normally the engineering resource needed to do this doesn’t match up with the gains in effectiveness and characteristic set by the end of the physical exercise.
#2 Wrapper bloat
Imagining about wrapper bloat: builders will generally build wrappers all over an current API or CLI from a vendor.
They will then incorporate far more characteristics, far more wrappers… and so on.
This can generate an unstable codebase that boils down to a copy of the first CLI or Rest API, but with the supplemental complexity of multiple wrappers to navigate.
#3 Fall the (co)pilot
Be cautious of co-piloted software program improvement.
In a way this is very little new, builders have been copying and pasting from Stack Overflow for many years.
The possibility with new equipment like ChatGPT is that there is no paper trail as to wherever the code arrived from. It is uncomplicated to envision a circumstance where it’s pretty much not possible to get to the bottom of an challenge with a substantial and sophisticated entire body of auto-generated code with no just one possessing the deep expertise of that code foundation that you can only get from creating it up slowly by hand.
#4 Can vs. should?
Let us communicate about the can you vs. should you conundrum.
It all boils down to the vintage concern of can you versus need to you i.e. automation offers software package builders substantial prospects to strengthen their workflows and increase new functions.
But it is all-far too simple to get carried away and introduce issues afterwards down the line.
CWDN Note: Akamai hosts a hard-core software program software improvement hub and web site web site that is easily searchable on the web exactly where Vereecke and other people go over command line issues from the cloud to the edge and everywhere in involving.