Saturday, October 08, 2016

Is software getting worse?

An interesting piece in The Register asks Is Apple's Software Getting Worse Or What?

Comment For over a year, Apple's software has been the subject of more derision than might be expected for a company of its size.

Developer Marco Arment took Apple to task early last year, arguing that OS X (recently rebranded macOS) is full of embarrassing bugs and that the company is trying to do too much on unrealistic deadlines.

Arment subsequently disavowed his post because of the widespread media attention it received. But there was blood in the water and the feeding frenzy has continued at Apple's expense, at least in part because controversy, manufactured or not, drives online traffic.

It continues to this day. On Tuesday, one fiction writer – who asked us to keep him anonymous – voiced his dissatisfaction, eliciting agreement from a few others. "I just need things that work, and that I can rely on working," he lamented. "I say this with the utmost regret, sadness, and no small sense of betrayal: Apple doesn't seem to make those things anymore."


The comments suggest it isn't only Apple churning out buggy software in the rush to add bells, whistles and intrusive data-trawling within excessively tight timescales. How many users want the bells and whistles anyway? 

"I just need things that work, and that I can rely on working". So do I and on the whole we get it, but have we reached peak software utility for home users? One comment which chimes with me is this.

little to do with apple

The fail fast fix fast mentality of software development is insane. (Have worked with software dev teams for 16 years now). Sounds fine if you are working on some new thing. But should not be used on core products. Whether it is apple (not a customer so can't say from personal experience ), Microsoft struggling with their updates, MANY others as well.

The focus has been shifting towards faster delivery of lower quality stuff because they believe they can just fix it later. Though in many cases later never comes because they move onto something else new and shiny.

It is possible of course to release things often but it requires more care than just doing it.

Too often agile is used as an excuse to ship faster and not need quality control.

Windows 10 seems to be turning into the largest scale agile fail in the history of software.

Companies like apple and MS have absolutely no excuses each having 10s of billions of dollars in the bank.

2 comments:

Nick Drew said...

aftr working 20 years in 'hard' industry (and being confident that whatever I sold would actually work) I took a 4-year diversion into software

our product was pretty good, but the pressure to ship new versions and upgrades was considerable, and the ability to maintain good QA ever-diminishing as ever more millions of lines of code were involved

we had an excellent CEO who demanded tight disciplines (never let the salesforce promise something we couldn't deliver - "if it ain't on the list, it don't exist"; always have contracts stating that the product is sold as tested; "we make sell excellent software, not perfect software"; etc etc)

but there were always bugs & glitches (not all of them out fault ...) and always dissatisfied customers to be pacified, the whole thing was pervaded by an underlying odour of permanent unsatisfactoriness

I only ever knew one absolutely pedantic and precise coder (a true craftsman - see your item on China!) who would never let his code out until it really, really worked - but he was viewed as hopelessly slow, and was let go

software is a very odd sector: I grew to hate fronting-up for product defects & all the other nonsenses, whoever's fault it really was: and I moved back to where I could be sure of satisfying customers

A K Haart said...

Nick - I well know that underlying odour of permanent unsatisfactoriness having drifted into lab IT. Our best systems were the very early, very simple standalone applications developed in-house. As the demands increased we moved over to commercial lab applications and things went downhill from there. Too complicated - that was the key problem.