Member Since: 16th Jul 2009
18th Aug 2010
The solution is obvious....
....based on their vast experience of this PAYE fiasco they can start again with a new project, because all the evidence shows a new HMRC IT system will save money and give a better service.
I hope everyone will be adding this thought to their response to the consultation for a proposal by HMRC to completely screw over PAYE.
5th Aug 2010
Many concepts bandied about but few definitions nor any sense of
"Real Time" sounds like one of those motherhood and apple concepts that is impossible to argue against (after all, who would argue for things happening in unreal time?)
But does anyone using the term (and let's not exclude HMRC) actually know what it means? Has it actually been defined anywhere?
Do they mean
(2) "sufficiently quickly so by the next time you look it's happened"?
(3) "asynchronously triggered"?
Any system has got to have a loop response time less than the its required update response time (otherwise it develops the equivalent of worklists and starts correcting its corrections)
Systems with short response times trade stability and reliability for flexibility and agility. They also cost more to build and to run.
(so, fighter aircraft can turn on a sixpence but can't glide for toffees, civil aircraft have a huge turning circle but can glide under no power for thirty seconds - and helicopters fall out of the sky when you take your hands off the controls)
The reasoning that a PAYE system needs to behave more like a fighter aircraft and less like an oil tanker because there are more people with multiple employments needs closer scrutiny.
Some people with multiple employments might have relatively stable circumstances so once the existing system settles down (PAYE codes have been sorted out) then all is well
Some people with multiple employments might have a major job and one or two minor jobs (e.g., 80:10:10) in which case they might be happy with a PAYE system that lags a little providing it eventually determines.
I am sure that the accountants among you will be able to develop other use cases.
So how many people (as a percentage) would see their PAYE performance significantly improve compared to a do nothing/ continuous improvement model.
Of course if one could build a single system that provides fighter aircraft capability with civil aircraft capacity (but let's not forget "reliable" "cost effective" "soft-fail stability" "hot-switch back-up systems".
Do we have any evidence that any government system has been designed with this "reliable" "cost effective" "soft-fail stability" in mind? (funnily enough, yes, the PAYE system as currenlty operated)
The government designed a instantaneous update systems for MOT - here's a reminder as to what happened:
it created chaos: http://www.theregister.co.uk/2006/05/16/mot_computer_crash/
MOTs are one thing, a system that collects 80% of the tax take is another.
Even if such a scheme were piloted, the evidence would not provides a basis for scalability, because a system changes in character as it scales (by way of analogy, the time it takes to get from one side of Waterloo Station to the other is not dependent on how fast you can walk, but the time of day you are attempting the task).
If I were designing a PAYE system it would be on big society principles rather than big government, I'd publishing the rule set, set minimum standards for participants, encourage multiple centres doing distributed processing, improve data logging, add store-and-forward capacity to manage variable data volumes and leave HMRC to collect aggregated data. There'd probably need to be a balancing mechanism for difficult cases.
What do accountants and payroll bureaux think about that idea? Any takers?
25th Jan 2010
I have deen trying to access SA Online all weekend and most of this morning. I keep getting a "System error" with a suggestion that I contact the help desk quoting a 16 letter reference.
Moreover I swallowed my pride and borrowed an MS Windows machine on the assumption that soemhow Linux had been disabled.
Anyone experiencing similar?
13th Aug 2009
Consider alternative software
Changing anything is obviously a PITA, but as information is the new money, using software that has an intrinsically better security model has got to be a good idea.
And there is no doubt that the better model is the one used by GNU/Linux.
Of course, if you leave all the doors open it's easy to walk in, but with GNU/Linux you've got to try very hard to leave the doors open. And there is no software in the world that will protect you from claiming your winnings in a Spanish lottery. Nor can it protect you from downloading the contents of a compressed file from a dodgy server.
However, and contrary to official advice talking about "all software", GNU/Linux saves you from yourself when you click on those dubious links. You can't accidentally or unknowingly install anything.
And anything you do download (I thought it was "dark side of the moon" not...) doesn't end up in some unknown and hidden folder somewhere.
I am aware that not all of your favourite software has its FOSS equivalents, and unfortunately, its the accounting packages, but take a look here, (it's a UN publication and it's already out of date, but it's indicative) http://www.apdip.net/publications/fosseprimers/foss-edu.pdf
But if the industry body cannot persuade the applications vendors to develop versions for GNU/Linux, surely it's big enough and ugly enough to fund the development of new or improve existing accounting applications?
One model for this development is google summer of code http://code.google.com/soc/ delete google insert accountancy bodies.
Usability studies have long since shown that changing from what you currently use to a desktop based on GNU/Linux is no more difficult than overcoming the problems encountered when one of your current suppliers upgrades an existing package.
The desktops are increasing pretty, for example: http://arstechnica.com/open-source/reviews/2009/08/hands-on-kde-43-deliv...
Everyone knows the servers are secure.
No it's not a pain free decision, but the pain is a one off experience. And then, if you use FOSS, no one will be auditing you to check for software licence compliance, no one cares if you use another copy of anything you use, no one will ask you for any money, either initially or for upgrades. It doesn't need such powerful hardware to run.
Of course you could choose to pay for support, but then you might choose to pay to have a washer changed on your tap.