Re: from a CDK customer
There's two kinds of 'good' support:
One, the good kind - where a product is well supported.
Two, the bad kind ,where the product is so shit the support are well versed with the various fuck ups.
862 publicly visible posts • joined 20 Apr 2010
Well, at its base, business is about money.
Saying outsurcing/offshoring is about money is not enough.
However ....
If you make a casr that outsourcing/offhsoring exposes the business to operational risk or, worse, large fianncial costs, then you are on the right line.
Me peroanlly, utsorucing/offhsoring would need to save me a lot of money and cost save me 50% before Id consider it worthwhile.
As it stands, each time Ie gone thru the initial figures, the saving from outsroucing have been pretty marginal - less than 20%. And the riskspretty high.
Whne I've gone thru actual figures, Ive found the outsourcing of my kinda work is actually more expensive.
Hmm.
Hey kids! Lets put a record o nthe jukebox and dance on the table ...
The issue with GDS and its 'wunderkinds' is that a bunch of Web front end 'pogrammers' are just that - a bunch of people who are contantly dicking around wrtiing and re-writing javascript libaries.
Total waste of time.
What you need is a stripped down webgui front end, customed to whatever service, an encryped trasnports and a tranactional backend.
And then those services have to be released, tested and updated in a controlled, managed way.
The bulk of the that work - the tranport and the tranaactional backend - are mind numbingly hard and skills you are unlikey to find in a shoreditch cereal bar. Or a bunch of Guardian reading HTML under 25yo bullshitters.
Slightly more complex.
Official java will bleed you dry.
Theres the open java thang - icedtea or whatever.
But you are on your own. Last time i looked oracle did not ship the regression test.
If youve made a sigificant investment in java then youve 3 choices
1- Be bled dry by oracle. Youll pay each time larry has a shit.
2 - Invest in your own java development team - produce your own regression tests for jdk and your apps. A massive commitment because if java is anything its a complex mess.
3- port everything away.
Pick one now.
Theres a lesson here.
Java was always a solution lookign or a problem, so the ammoutn of code that went into it was shocking - 4-5 guis, various means of networking, god knows how mnyn 'Wouldnt it be nice ...'
All over a 20 years period. And the they were reluctant to deprecate and delete the old crap.
*IF* Java was a small, itght VM with support for sockets and decent built in integer then it would rule business and web stuff. No brainer.
Instead is a bloated monster which requires a lot of money to just get the basica changes QA in place.
Any compnay that bet on jave - and a lot did, internally - are o nthe hook for ongoing support costs. And, to put the cherry on the shit cake, that person is Oracle.
Me? I stick C99 and its standard runtime. Nice n small, with lots of competing suppliers.
Oh that £9m is mine.
Maintain a list of people: NHS trust CEO, heath secretary and handul of MoH senior servants.
Sack them all if there's a data breach.
This prize bullsht is annoyig. It works well for moonshot stuff -reuable rockets where there's various billionaires funding the development.
It doesnt work for the everyday, should be doing this default stuff.
I smell a bright young PPE/Ecomomics Oxbridgey civil servant's hand in this.
Ouch.
Id hate to be holding equity in a bank thats has mesed up after, say, 2016.
Just as well as all the UK banks are totally robust and haven't spent he last 6 months in various states of fuckup.
~32m, reduced to 16m for cooperating.
Id be putting those number int othe outsourcing cost saving spreadsheet.
No.
You are making the classic mistake of saying 'cant do IT'.
The assumption is that they are OK at the not IT stuff. They are not.
The problem is that the NHS, UKGOV/public sector do no purge people. The private sector as recession and competition - adapt of die. The private sector is not good at IT/software because it wants to spend the blood sweat and tears. They are good because they HAVE to be - or die. See TSB for this.
The problem with UKGOV/NHS/public sector is people get their feet under the desk and move up according to age/brown nosing. Anything disruptive/competitive o that position gets pushed back on.
For a bureaucrat, theses nothing more undermining than software - poof, jobs gone in flash.
If you want to get the same level of change/adaption in the NHS/UKGOV/public sector then you need to getting rid of a good 50% of the none clinical/techinical positions every 10 years - spoof recessions. Do it by random selection otherwise the same self import clique will remain.
And Im serious on that.
People who work in healthcare, esp, management, are self important, process and personal responsibility avoiding morons.
Not sure why. Its chronic.
And before you say - 'Ah its only computers/software ..' they apply the same level of competence and ability to clinical stuff too.
Nim.
Another attempt at avoiding the issue - javascript is shit.
Some solutions:
- Fix the fucked up bits of java script. Stuff like Typescript are a convoluted way of doing that. But looking at the many versions as it stumbles around generating work for the MS typescript team, is very unappealing to me.
- Drop Type script and use C, stripped Python .. anything but really that is not too complex.
- Standardise WebASM so it does nt matter hat language you use.
No.
300 column lines apart, the less code, the clearer the ptogram.
CAR Hoare
'There are two ways of constructing a software design: One way is to make it so simple that there are obviously no deficiencies, and the other way is to make it so complicated that there are no obvious deficiencies. The first method is far more difficult. '
Its one reason why Ansi C is successful - simple lannguage, simple libraries.
My issue with complex languages and complex libraries is not whether i can use them - I can! Trust me!!
Its whether the 3rd party delivering and maintaining them doesnt fuckup on some version down the line.
The success of any long running software products depends on hiw easy the product is to test and update.
The logic describes what a program does.
The syntax describes how it does it.
Java's syntax is way too verbose so the wheat (logic) gets lost in the chaff (syntax).
It takes years to see thru syntax and see the logic.
Whats missing from Java is a decent run-time, something to join up programs.
Objective_C has one and its proved very successful.
The closest Java comes is a hulking server.
Oh, and being able to easily and safely split a java program into libraries. 20 odd years they are solving that. Way too slow.
To quote Joe Armstrong - If you are writing code for a DV player then Java is fine.
Designing processors and instruction sets is hard.
Sure, the first steps go quickly - oh this instruction willl do this.
Then you throw stuff out, put stuff in.
You corrupt the nice simple isa with special cases.
That chesper n faster chip turns out slower and more expensive. And investors want something now!
The elephant in the room is a physics.
Intel get where it is by making a crap design run on smaller and denser transistors.
The capital required to get much smaller is no longer economic, at least for a horizontal integrated company, making and selling its own chips.
Intel will be wiped out by a company that cracks large scale cores, large caches, running at 3Ghzish.