Microsoft whips out PowerApps – now your Pointy Haired Boss can write software, too! Microsoft has announced PowerApps, a new way to create and host applications for its Azure cloud service. PowerApps is an "enterprise service for innovators everywhere to connect, create and share business apps," says Application Platform VP Bill Staples. Sure, but what is this really? Microsoft, it turns out, is still …
Access is (was) supposed to be a straightforward way to build apps for data without needing to know code, but every one I've ever encountered built by non-developers has been something of a hash job, needing a knowledgeable developer to get things working right. I kind of expect PowerApps will end up the same way - lots of half-baked attempts needing reworking, then finally being junked in favour of something more robust and usable. The only 'advantage' I can see is that unlike Access with its proprietary hidden code, the interface system is built on known open standards like HTML and Javascript, etc, which might make bits more capable of being reused.
<quote>Sure, if you are a FuseBox developer - make no mistake, before IT managers Executive Damagers have an idea about the difference between Java and JavaScript there must be a half-a-million Billion years pass.</quote>
FTFY!!!
On other topics, I just graduated to a silver badge. How nice!!!
"Access is (was) supposed to be a straightforward way to build apps for data without needing to know code"
But as it was built on an RDBMS unless its users needed to know how to design a database. If they couldn't get that right no amount of code knowledge avoidance was going to help.
My recollection from building a couple of little applications in it was that Access (when I did it, Access 95 probably, if such a thing existed (CBATG)) seemed a bizarre and cruel joke to me in that it was harder to build fairly simple but non trivial form based applications in it than it would be in Visual Basic.
Software-Thru-Pictures (an expensive one that the DoD really liked for just a little while)
I actually don't want to do an inventory of these things that will keep the PHB busy and out of the "Software Engineers" hair.
However SW development shouldn't be so damn fragile and its practitioners shouldn't feel like the shamens of the universe.
Someday, someone will figure out a plug-in system where I can take my peripheral (a lamp?) and stick it into a socket to get power. Or a common communications bus. Or a common way to write specifications.
And there was me thinking I was the only one who could remember that piece of 1980s crud. It wasn't the only one either, the software house I worked for wrote its own after we found out what happened when you clicked the button that said "Don't click this button" or was it a menu option.
Have an upvote.
Might be useful, in the past I've worked places where Access was used to develop something ad hoc which was then used as the basis for requirements etc. if it turned out to be useful. It would most probably be called an agile approach these days. The trick as always is not to allow it to get out of hand.
Might be useful, in the past I've worked places where Access was used to develop something ad hoc which was then used as the basis for requirements etc
The problem is that "ad hoc" often has a nasty habit of becoming "leave it running, it's your job to keep it up". This is why quite a few people I know reverted to the use of mock-up screens only - pure self defence...
Yes, this used to be called Rapid Application Prototyping / Rapid Appication Development and has been re-invented as Aglie. By designing the logic flow with/by the users, not bothering with any specialist IT skills, limited testing, negligible documention, absolutely no version control, systems management or handover an application can be brought into production very quickly and inexpensively, then maintained at great effort, risk and cost for all time thereafter. Dont get me wrong RAD/RAP/Agile have their place, but its a double-edged sword.
Actually, this looks great. The PHBs can build and maintain the shit they want - that generates the meaningless pie charts they love from the unreliable and unverifiable data they demand - and leave the actual important work to the people who know what they're doing. I can see a future where businesses run in a completely smooth two-tier system, where management generate their own software to produce the pointless reports they need to make themselves look good while the people who do the actual work are left to get on with it unhindered. I might actually be tempted back into the industry.
"The problem is that when the PHBs start churning out a few of these easily they will start to question why it takes you so long to get an application delivered."
Already happened to me. A few years before I retired a new PHB entered the company and informed me on day 1 that he didn't see why it took us so long to develop reports, he had seen Microsoft demo a tool that just allowed you to drag and drop data from a database and display it on a website in a couple of minutes.
(A year later he left when the app for which he had been managing the development turned out to have no business logic behind it.)
It may make applications easy to build but they are going to be limited in what they can do and very inefficient. Then the help desk is going to get a bunch of calls wondering why the computer and/or network is slow when the boss is running a bunch of these.
Mind you developers can be very inefficient too. I once had to fix a web application that would take minutes to load because the developer decided to use an O(n^2) algorithm in JavaScript on the browser to combine two sets of data rather than just doing a join in the database.
"Mind you developers can be very inefficient too."
Mmmm. I remember the program that was going to take over 24 hours to load a day's data into the replenishment system. And that was the second effort of that vendor's I had to debug for them. The first just asked the database engine to spawn more & more objects without freeing any or re-using them until it burst.
"I remember the program that was going to take over 24 hours to load a day's data into the replenishment system."
Or, in mainframe batch terms:
"Daily runs for a day, weekly runs for a week, and no one has ever completed an end-of-quarter."
This scenario really happened during a development project back in ~1985.
Yet again a tool to solve the problem of coding being "complicated", which it isn't, to make money from the folks who think it hand this shiny new magic bullet will let them write code magically to solve any problem. That reality of automated development just got even further into the future, since we'll get another decade wasted messing around with this and fixing random dross.
Back in the day - <voice shaking> - we used to call these things 4GLs - (Fourth Generation Languages for those who grew up on Microsoft) - they just created crap code and added very little value. Anyone that could code in C could generate much better code and results.
Can't see these new fangled code generators been much better. Of course if it uses up lots of CPU cycles, network traffic and storage, your average Very Large Cloud Vendor is not going to be displeased to send you a large bill or as they call it "Creating demand"
C'mon Lancelot - time to search for the Holy Grail again
Actually both the examples I gave in another post were coded in 4GL. But that wasn't the problem - exactly the same mistakes could have been made in C. The problem was failing to understand how to use the database engine (which was, of course, written in C). An efficient language is no obstacle to writing inefficient code.
Just what I was going to mention, actually it was:
70s: CASE Tools - going to make programmers redundant any day soon.... (where is IEF or coolgen as it later became these days?)
80s: 4GLs dBase/Foxpro et al so 'power users' can build apps - which sort of worked
90s: RAD tools, anyone (and I mean anyone) could walk into PC world, buy a copy of VB3 and label themselves a 'programmer' and then hash up fucking nightmare apps
00s: code generators and 'scaffolding' gets you somewhere as long as that somewhere looks and acts the same way
Its bad enough seeing most businesses run off a million and one excel spreadsheets which have become 'apps', still it means instead of steaming piles of crap all over the place its all going to be in 'the cloud'
The essential problem isn't that coding is hard; it's that requirements are hard and expressing logic is hard, and mapping requirements to logic is really hard.
We've seen this movie many times before, and it always goes the same way:
Act 1: Our hero releases a tool to let anybody code. People are excited.
Act 2: People start to ask for more capability in the product, because they can't express the things they need to do / reuse the function they created once / find that thing that needs to change because the business rules changed. Our hero reluctantly complies, although he is concerned that the tool is getting big and complex.
Act 3: Our hero wakes up one morning to discover that his tool has become every bit as unwieldy and difficult as a full-blown development environment... because it is.
Coding is mental mapping, real-world challenges to real-world solutions, and when that is done by more than a few data tables and HTML transacts it becomes really exhausting. We are not idiots or stupid - we are just really, really tired. Thanks for posting and making this somewhat obvious!
... isn't this what Microsoft, intel, AMD, Apple, google, ... have been doing for years?
Making IT and digital stuff doable, accessible and workable to the greater populace out there?
No?
EDIT: plus it reads more like Windows PowerToys than Excel macros (or even more like .bat workery)
I wish it was doable.
I admit, it might be profitable to some, but be honest - just how many people are truly enticed to install applications on their micro-powered phone (1MB to spare?!)
Pathetic, and the android space is just as bad if not worse.
I want 1 TB RAM - Period.
I want 1 Petabyte Storage - Period.
Nothing less on any platform by my standards measure up.
Have a great day!
PowerApps is an "enterprise service for innovators everywhere to connect, create and share business apps,"
There's the hook... PHB's and their ilk all believe they are "innovators" because the corporate rhetoric says they are. This always translates into a nightmare for us in the support areas that have to support their the junk they come up with. The junk level is now being taken to a new low or to a new high depending whether you're digging a hole to bury it in or shoveling crap on top of it.
The entire idea that there "aren't enough developers" is horse manure. There are a functionally unlimited supply, many of them trained, experienced and idle. The problem is that nobody wants to pay them a living wage or provide humane working conditions. So the developers leave the field and go elsewhere.
The new developers brought in to replace them aren't as experienced. They are willing to work for peanuts at first, but realize quickly that living with 12 room mates in a 3 bedroom that's 400sq ft is asstastic and they go do something more profitable.
If you want more quality coders pay their rates. Otherwise you get commodity software developed by commodity developers.
The apparent quote "aren't enough developers" isn't actually a quote from the story; the story actually says "aren't enough skilled developers".
In my experience there are plenty of insufficiently skilled developers (many of them "trained, experienced and idle" to quote yourself) but there is most definitely a skill shortage in the industry despite good pay rates. I have personally never met a truly talented IT person of the type I recruit (developer or otherwise) who left the IT field to do something else more profitable.
"I have personally never met a truly talented IT person of the type I recruit (developer or otherwise) who left the IT field to do something else more profitable."
Then you haven't met a lot of people. I can point you at well over a thousand top-flight nerds who've left the field for something more profitable. Most of them into management. Some into jobs that are only equally as profitable but where the hours are far more flexible and they have way less stress. (This was a route I took, though I would be among the least qualified of the group I am talking about.)
There comes a point where $200k is just not worth the misery that comes with the job. Especially when you can get $200k just about anywhere else for far less effort.
Skilled developers are a dime a dozen. And if I am perfectly honest, the skilled developers never really stop being developers: they continue writing programs all the time, but they stop doing it for corporate overlords.
Those developers might become management or marketing or what-have-you, but they then contribute to open source, or write things that amuse them or they transition to infosec, using their skills for hacking.
Ahhh, infosec. Infosec absorbs good developers like a black hole accreting matter. If you are, in fact, a skilled developer you can take you $100k salary and make it $250k over night. A $250k developer can be a $1M infosec nerd in few months.
So sorry, mate. If you honestly think that skilled developers - or even mediocre ones! - stick with the punitive position of "corporate bitch" for overlong, we have some very different definitions of skilled.
By definition skilled developers are not 'a dime a dozen' because skilled refers to the upper tier of developers, which is by definition less than 50% of developers.
Additionally, a developer who gets promoted to manager isn't leaving the field.
And I refuse to believe developers have left to go into marketing. Completely different skill set that is diametrically opposed to programming. Developers are the ones who spend time removing ads and spam from their daily lives, not people who gleefully generate it.
"Good pay doesn't generate experienced developers"
Experience generates a demand from developers for good pay.
Treating them like shit and not meeting their pay demands encourages them to seek employment elsewhere.
But please, go run your little company in the belief that you can always win with the worst paid people. I'll enjoy your bankruptcy.
> The new developers brought in to replace them aren't as experienced. They are willing to work for peanuts at first, but realize quickly that living with 12 room mates in a 3 bedroom that's 400sq ft is asstastic and they go do something more profitable.
Simple: don't live on the west coast. Plenty of programming jobs elsewhere without the stupidity of California.
"...According to the documentation:
The message contains a link that a user can select to access the app. Any users who don't have PowerApps (or aren’t signed up to use it) are prompted to install it and sign up for it..."
So... Microsoft documents with built-in scripting capabilities, distributed by email. I'm pretty sure that's been thoroughly thought through and nothing could possibly go wrong. I mean, it's not like that's ever been used as an attack vector, ever before. Is it?
As in "can you build a Trojan in PowerApps?" Probably not.
I would tend to agree that sending emailed links is a bad practise and one we should discourage because it trains users to click on them but they already do and nothing seems to change it.
This post has been deleted by its author
Can be fairly safely said that configurable stuff probably has its merits. But allowing vast legions of dim eejits to "play programmer/analyst" is just f*****g stupid. PHB-eejits ... worse yet!
While the stupidity is nothing new (brilliant reference to The Last One, amongst others), the Emperor's New Clothes* manage to take the pointy-clicky-shiny-stupidity to a new level.
/rant over.
It's beer o'clock on a Friday somewhere, Somewhere it is so ...
*PowerApps
I realize that this is a matter of opinion but I think that Microsoft has some pretty impressive and useful technologies in their portfolio. For example; a full VBA engine underneath their Office environment; it combines the easily accessible BASIC language and combines it with a pretty slick OO based model allowing you pretty much to glue your Office needs together.
Microsoft also has a problem: usually when they start out with something then it's not always state of the art. I don't think I have to come up with examples; we all know our own Microsoft horror stories. However, it's also fair to say that Microsoft has shown the drive and the skills to turn things around. Drastically. Turning something utter shite into something very useful. I know, I know: matter of opinion.
But here's the thing I fail to get: once they managed to reach that point then they usually abandon their stuff all together. Example? Well, how about Expression Web and Expression Design. An environment made for website development combined with vector graphics support. People familiar with the Dreamweaver / Fireworks combination would feel right at home. I've used Expression Web 4 professional myself, even paid for it, and it was very useful. It allowed me solid HTML design but also provided a bit of programming features.
And there came Visual Studio 2012, which had pretty horrendous start itself. Yet it was still determined that it should replace Expression Web effectively replacing a tool fully aimed at web development with a tool aimed at general development. It was horrid at first, but they did come around. My point: why throw away a working environment in exchange for something not even complete?
You see this all over the place. .NET anyone? For years they pushed .NET forward as the de-facto standard for programming on Windows. Then we got the mobility fever and wham: it had to be replaced / enhanced with a Javascript engine out of all things. Something else? Windows 7: an extremely popular operating system, for many a true replacement for XP. Instead of pushing it to its limits: no, we're going to re-invent the wheel with Windows 8 (disaster) and then Windows 10 (which still has to proof itself).
The problem is that they're fracturing their own market up to such extremes that it's hard to keep up, let alone making sure that you really want to invest in whatever is hip right now because chances are always high that it can get dumped and replaced with something else anytime.
And here is yet another example... There was Frontpage, which I agree wasn't exactly very good. Then we got Sharepoint which showed some pretty serious potential, still had its quirks of course but it mixed pretty good into their Office suite. And so here we are: lets dump it and move on. Even though Sharepoint has come such a long way and is a pretty solid environment right now.
Why can't they stick with their own program for a change and back it up with everything they have?
I haven't done a full comparison of the available features, but the "send email when there's a new tweet" thing sounds awfully like what one can already do (totally for free) on ifttt.com ("If this, then that"). That site also has a great many recipes that others have already written if one wants to copy something to get started.
Very few users have worked with PowerApps at this point so we don’t know how easy it will be to use, nor do we know how robust the applications may be. There are already critics suggesting that the created apps will be very substandard compared to most, although much of that criticism may be coming from programmers attempting to preserve their own role. I suspect that while the apps may not be as robust as some, there is a huge market for a tool to enable end users to capture and manipulate relevant data for their own needs without having to rely upon a programmer to write the code. While it is possible to share the apps with others, the goal is probably not one of creating applications for distribution but rather, creating custom apps for individual use.
It is early days, however we have seen many past attempts at this.
- COBOL
- "The Last One"
- dBase
- VisualBASIC for Applications
Writing code isn't the hard bit. That's surprisingly easy. The difficult bit is understanding that computers are dumb, need everything spelled out in minute detail, and will not stop and think: "Hey, this doesn't look right!"
They'll just blunder along as programmed, regardless of the intent of the software author.
The hardest bit to deal with is changing requirements. One can write a program to perform a function, and have it working fine. Then the goal posts move. No matter how good a coder you are, you will wind up making some assumptions. Sometimes you have to hard-code them, sometimes you can make them configurable.
I've been bitten many times by making an assumption at one point, only to find a use case that breaks my model later on. You can usually only avoid this problem in the most trivial cases.
COBOL actually worked, because it enabled people who had never been programmers before (in a world which until then required one to be an A M Turing or a Max Newman to get computers to do anything) to develop programs some of which are still doing mission critical stuff. That's why Grace Hopper is part of the computing pantheon.
Everything since then has arisen because of the desire of the proponents of MBSA (Management by shouting at) to leave unsaid things that need to be said - as you say.
COBOL indeed worked, but it wasn't without its problems that needed a skilled programmer. The problems COBOL faced aren't that different to what's faced today in JavaScript and C# for example.
You still needed clear requirements. Those requirements would be based on certain assumptions. When those assumptions no longer held true, the program broke and needed fixing.
Without careful consideration, that code can soon grow into an impenetrable tangled forest. Your only option then is to slash and burn.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
