SERVING THE QUANTITATIVE FINANCE COMMUNITY

Posts: 23951
Joined: September 20th, 2002, 8:30 pm

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

The PC sales continue to slump -- down again in 2017.

iPhones now outsell desktop & workstation PCs more than 2:1. (216 million vs. 98 million)

And smartphones now outsell all types of PCs more than 3.6:1 (1.5 billion vs. 423 million)

alanW
Posts: 3
Joined: March 1st, 2018, 7:33 pm

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

I was told FinTech could change the market, such as http://www.finpricing.com

Cuchulainn
Topic Author
Posts: 62387
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

I was told FinTech could change the market, such as http://www.finpricing.com
You're all over the place, old bean!

Cuchulainn
Topic Author
Posts: 62387
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

The PC sales continue to slump -- down again in 2017.

iPhones now outsell desktop & workstation PCs more than 2:1. (216 million vs. 98 million)

And smartphones now outsell all types of PCs more than 3.6:1 (1.5 billion vs. 423 million)
Interesting.

P.S. in the spirit of the thread, what is the great FIVE year plan smarties versus PC?

Posts: 23951
Joined: September 20th, 2002, 8:30 pm

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

The PC sales continue to slump -- down again in 2017.

iPhones now outsell desktop & workstation PCs more than 2:1. (216 million vs. 98 million)

And smartphones now outsell all types of PCs more than 3.6:1 (1.5 billion vs. 423 million)
Interesting.

P.S. in the spirit of the thread, what is the great FIVE year plan smarties versus PC?
Projections seem to suggest continued modest declines in PC sales. What's interesting is that smartphone sales seems to be slowing, too, although other data shows that people are holding on to older models much longer.

Overall, based on the ratio of smartphone-to-PC sales and ratio of smartphone-to-PC longevity, it looks like the hardware landscape will stabilize at maybe a 2.5:1 install base ratio.

Cuchulainn
Topic Author
Posts: 62387
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

The PC sales continue to slump -- down again in 2017.

iPhones now outsell desktop & workstation PCs more than 2:1. (216 million vs. 98 million)

And smartphones now outsell all types of PCs more than 3.6:1 (1.5 billion vs. 423 million)
Interesting.

P.S. in the spirit of the thread, what is the great FIVE year plan smarties versus PC?
Projections seem to suggest continued modest declines in PC sales.  What's interesting is that smartphone sales seems to be slowing, too, although other data shows that people are holding on to older models much longer.

Overall, based on the ratio of smartphone-to-PC sales and ratio of smartphone-to-PC longevity, it looks like the hardware landscape will stabilize at maybe a 2.5:1 install base ratio.
So, it's plateauing? What's next for the hungry consumer?

Collector
Posts: 4647
Joined: August 21st, 2001, 12:37 pm

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

5 years from now? uncertanity has collapsed, encryption systems no longer work! How would that affect computing?

And besides prediction above, not exactly just software, but close enough:
Smart clocks should be much more useful for other things than just fun" by then, so expect growth in smart clocks. Todays smart clocks looks mostly ugly, and has limited functionality, future elegant models where one can solve math problems, price complex options and do physics is in demand. Some kind of zoom out screen would be nice, but hard to develop for decent price and also to keep robust.

Surge in robot-technology.

$\approx$ 2060 = Idiocracy  (Newton predicted that the world would end no earlier than 2060)

Cuchulainn
Topic Author
Posts: 62387
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

Prediction
"When developers learn to design parallel programs it will be done using functional programming models."

Posts: 23951
Joined: September 20th, 2002, 8:30 pm

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

Prediction
"When developers learn to design parallel programs it will be done using functional programming models."
I think this is true but the causal pathway is different than the one stated.

The deeper problem for "good" parallel design is that most humans are incapable of parallel thought (actually, most humans are incapable of algorithmic thought and only a fraction of algorithmic thinks can think in parallel algorithms). Most developers are pretty much unable to get beyond thinking in loops because they innately convert parallel activity into a linear sequence (and make tons of mistakes as a result).

It's not some much as matter of developers learning this stuff as filtering out would be developers who can't get beyond the loop-based world.

Cuchulainn
Topic Author
Posts: 62387
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

On a deeper level, the mental blocks are

1. Think about WHAT before HOW
2. Can you explain the DATA FLOW in 5 minutes to the customer?
Last edited by Cuchulainn on March 4th, 2018, 7:06 pm, edited 1 time in total.

Posts: 23951
Joined: September 20th, 2002, 8:30 pm

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

On a deeper level, the mental blocks are

1. Think about WHAT before HOW
2. Can you explain the DATA FLOW in 5 minutes to the customer?
Quite so!

Yet even the "WHAT" is a problem to a person who thinks about sequential outputs rather than simultaneous outputs. Perhaps this is Sapir-Worf at it's worst because 100% of human verbal languages are sequential.

Why explain data flow to the customer? (Actually, now that I think about, do developers using functional programming languages even think about data flow which seems be an internal hidden process entirely managed by the compiler or environment???)

Cuchulainn
Topic Author
Posts: 62387
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

Why explain data flow to the customer?

In this way both customer and developer achieve a common understanding of the business process before jumping into code. The context diagram and high-level DFD are invaluable.

(Actually, now that I think about, do developers using functional programming languages even think about data flow which seems be an internal hidden process entirely managed by the compiler or environment???)

In general, NO.

The internal (implicit) mental model is not based on functional flow (it used to be until the late 80s just before OOP revolution but the baby was thrown out with the bath water).

It is all a communication/language problem. If Sapir-Whorf is viewed as a continuum in a given field, then it has little vocabularies sprinkled on it. A good example is explaining to a GUI developer with no MCAD experience how to design a UI to model  a complex oil rig in terms of whole-part and parts explosion composite structures. The distance between the two languages is so great and the two parties will be at loggerheads until the budget runs out. And you can't expect the engineer to understand the niceties of GUI dialog boxes.  A mediator is needed.

Everyone knows these kinds of issues if you have worked for more than 2 weeks in a software team.

//
No two languages are ever sufficiently similar to be considered as representing the same social reality. The worlds in which different societies live are distinct worlds, not merely the same world with different labels attached. Edward Sapir

The diversity of languages is not a diversity of signs and sounds but a diversity of views of the world.
Wilhelm von Humboldt
Last edited by Cuchulainn on March 5th, 2018, 8:55 am, edited 3 times in total.

Cuchulainn
Topic Author
Posts: 62387
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

Yet even the "WHAT" is a problem to a person who thinks about sequential outputs rather than simultaneous outputs. Perhaps this is Sapir-Worf at it's worst because 100% of human verbal languages are sequential.

Recently, task dependency graphs are becoming popular as a tool to model parallel systems. Two tasks in the graph without dependency can be run in parallel. C++11 and Python use futures to model and are much safer than threads.

It is too late IMO to parallelise things _after_ having written your (sequential) code. A follow-on is that existing legacy OOP code cannot be parallelised, really. But you might get lucky in places.

Posts: 23951
Joined: September 20th, 2002, 8:30 pm

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

Excellent points about CAD & GUI. That's a very interesting and complex area given all the different UI modality options: light-pen, tablet, mouse, touchscreen, haptics, function keypad, and keyboard. At some level, the UI designer must conform to the demands of the engineer. And yet innovations in UI can mean that sometimes the engineer benefits from conforming to the UI technology. Customer & code must co-evolve. Too many requirements documents hardcode obsolete practices.

The biggest Sapir-Worf issue here is that there are zero parallel verbal human languages. They are all sequential which makes discussions of requirements, data flows, tasks, etc. inherently expressed in sequential terms. As big an issue as differing vocabularies might seem, differences in baseline grammar and structure are much worse.

Talking to the customer is likely to create sequential pseudocode (i.e., hard to parallelize) with the added problem that the customer has no clue about the potential permutations in the process (inputs, intermediates, or outputs) that could conceivably arise (see Therac-25 for a lethal example of this). I wonder how many race conditions occur because neither the customer nor the programmer were even aware that there was second runner on the track.

Cuchulainn
Topic Author
Posts: 62387
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

### Re: FIVE years from now: can you predict the Software Landscape in anno 2023?

Parallelisation is an optimisation step; The 'customer' does not need to know this. But he does know what 'guaranteed response time' is.