SERVING THE QUANTITATIVE FINANCE COMMUNITY

 
User avatar
Cuchulainn
Posts: 56933
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: Anyone getting into Rust

April 27th, 2018, 2:50 pm

How does one collect said evidence without using the language enough to see what it can and cannot do?

Example: Microsoft benchmarked C++ and Java and came up with C#.
Which answers your question.
 
User avatar
Traden4Alpha
Posts: 23951
Joined: September 20th, 2002, 8:30 pm

Re: Anyone getting into Rust

April 27th, 2018, 3:15 pm

Cuchulainn wrote:
How does one collect said evidence without using the language enough to see what it can and cannot do?

Example: Microsoft benchmarked C++ and Java and came up with C#.
Which answers your question.
And yet they failed to avoid these problems: http://www.informit.com/articles/article.aspx?p=2425867
 
User avatar
Cuchulainn
Posts: 56933
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: Anyone getting into Rust

April 27th, 2018, 3:25 pm

Traden4Alpha wrote:
Cuchulainn wrote:
How does one collect said evidence without using the language enough to see what it can and cannot do?

Example: Microsoft benchmarked C++ and Java and came up with C#.
Which answers your question.

And yet they failed to avoid these problems: http://www.informit.com/articles/article.aspx?p=2425867

Thanks for the link. Very nerdy stuff. IMO they are inconsequental. They are syntax-driven blemishes at most.

I do not consider this to be evidence. A better example would be a pricing library.
 
User avatar
Traden4Alpha
Posts: 23951
Joined: September 20th, 2002, 8:30 pm

Re: Anyone getting into Rust

April 27th, 2018, 7:52 pm

Cuchulainn wrote:
Traden4Alpha wrote:
Cuchulainn wrote:
How does one collect said evidence without using the language enough to see what it can and cannot do?

Example: Microsoft benchmarked C++ and Java and came up with C#.
Which answers your question.

And yet they failed to avoid these problems: http://www.informit.com/articles/article.aspx?p=2425867

Thanks for the link. Very nerdy stuff. IMO they are inconsequental. They are syntax-driven blemishes at most..
That's probably true. And yet every syntax-driven blemish adds labor-hours to the development and bugs to the final code.

Cuchulainn wrote:
I do not consider this to be evidence. A better example would be a pricing library.
Hmm... A pricing library is even less an example of evidence of the quality of a language because it may also be evidence of: 1) the programmer's failings to understand pricing; 1) the programmer's inexperience with the language; 3) the reader's understanding/requirements for a pricing library; 4) the programmer's inexperience with the language. At best, given chunk of code written by a domain expert with good knowledge of the language and reviewed by a domain expert with the same library requirements and with good knowledge of the language is a single data point.
 
User avatar
Cuchulainn
Posts: 56933
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: Anyone getting into Rust

April 28th, 2018, 1:57 pm

This is not what  I mean, really.

These precondition are missing 5) and 6)

5) the programmer has incomplete knowledge of the mathematical/numerical methods and relies on recipes which may not work in the current context.(e.g. nonlinear solversa and the deluge of recent articles on tweaking them.)

6) programmer knows too much of the language (potential risk item is product becomes over-engineered). Let's face it, developers love new cool features.

My point (and the thesis in the u tube video) to examine the postconditions when the software prototype is delivered. Call it a post-mortem analysis whose deliverables are a checklist for improvement., similar to quality improvement in Japanese work processes. The conclusion could be to mitigate risk items 1)-6). It's a part of project management and should be flagged during project estimation.
 
User avatar
Cuchulainn
Posts: 56933
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: Anyone getting into Rust

April 28th, 2018, 2:06 pm

Barry Boehm's top 10

Top-Ten List of Software Risk Items

1. Personnel shortfalls
2. Unrealistics schedules and budgets
3. Developing the wrong software functions
4. Developing the wrong user interface
5. Gold plating
6. Continuing stream of requirement changes
7. Shortfalls in externally furnished components
8. Shortfalls in externally performed tasks
9. Real-time performance shortfalls
10. Straining computer-science capabilities


Happens all the time
Its going to happen - happen - till your change your mind
Its going to happen - happen - happens all the time
Its going to happen - happen - till your change your mind
 
User avatar
snufkin
Posts: 63
Joined: January 25th, 2017, 9:05 am
Location: Cambridge

Re: Anyone getting into Rust

August 3rd, 2018, 6:20 am

I'll just leave it here: https://twitter.com/AndreaPessino/status/1021532074153394176

It's finally happening - after >30 years of pro use, 20 of which quite reluctantly, I am officially DONE with C/C++. Only maintenance from now on, everything new will be in@rustlang. THANKS Rust team for refining modern concepts into such a practical, elegant system. I love it.
 
User avatar
Cuchulainn
Posts: 56933
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: Anyone getting into Rust

August 3rd, 2018, 3:52 pm

Programming languages are like religions; there are so many to choose from.

I am taking a more sceptical viewpoint because the longevity and desirability of a language in an industrial organisation depends on many factors:

Developers tend to choose languages based on cool, 'modern' syntax (90% of which is >40 years old, anyways) and the decision tends to be subjective, context-sensitive and idiosyncratic. Management has different views.

In the tweet, words like 'modern', 'elegant' and 'practical' signal the alarm bells start going off for me ... developers love beauty in the language. I see it with C++; as soon as you say 'perfect forwarding' they become excited.

I experimented a bit with Rust and it does look quite good, a bit like minimalist C. But only a first impression. It may be the future, who knows?

What are your selling points to potential disciples of the "Rust Evangelism Strikeforce". I could not find a 10-point feature list.  Hard to believe on just word-of-mouth.

I don't think C++ production will be ported to Rust  any time soon. HOWEVER, if you can call C and C++ easily from Rust then maybe.. It is a project estimation problem.

Rust is like gradient descent ... it is a local optimiser IMO. All languages are just a variation on C, ML, Fortran and Lisp. Nothing new under the sun. It's probably just another language. But time will tell.

But I think it is worthwhile to take a project in Rust. Good luck :)

P.S. My wife asks "What about waiting for Rust++?"
 
User avatar
Cuchulainn
Posts: 56933
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: Anyone getting into Rust

August 3rd, 2018, 5:16 pm

Looking into crystal balls ... Rust will never be used for computational finance. 
 
User avatar
snufkin
Posts: 63
Joined: January 25th, 2017, 9:05 am
Location: Cambridge

Re: Anyone getting into Rust

August 3rd, 2018, 11:53 pm

Cuchulainn wrote:
Looking into crystal balls ... Rust will never be used for computational finance. 

Sheesh, I have a few projects to try it right away. What do you mean by "used" here? What's your definition of "never"?
 
User avatar
snufkin
Posts: 63
Joined: January 25th, 2017, 9:05 am
Location: Cambridge

Re: Anyone getting into Rust

August 3rd, 2018, 11:58 pm

Even Scala is used for computational finance, so I won't be surprised by anything at all. Haskell is used, with all its "academic" background. People don't care that much these days: if it works, it's good enough.
 
User avatar
Cuchulainn
Posts: 56933
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: Anyone getting into Rust

August 4th, 2018, 10:18 am

snufkin wrote:
Even Scala is used for computational finance, so I won't be surprised by anything at all. Haskell is used, with all its "academic" background. People don't care that much these days: if it works, it's good enough.

I can see why Haskell and Scala are useful. 
But what are the features in Rust that makes it suitable in the current context? 
 
User avatar
Cuchulainn
Posts: 56933
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: Anyone getting into Rust

August 4th, 2018, 10:21 am

snufkin wrote:
Cuchulainn wrote:
Looking into crystal balls ... Rust will never be used for computational finance. 

Sheesh, I have a few projects to try it right away. What do you mean by "used" here? What's your definition of "never"?

Crystall balls don't always work. It's a guesstimate.
What kinds of project? What I'm asking is the compelling reasons for moving to Rust in this field?
In engineering one carries out a proof-of-concept (POC) project to test feasibility. 
 
ABOUT WILMOTT

PW by JB

Wilmott.com has been "Serving the Quantitative Finance Community" since 2001. Continued...


JOBS BOARD

JOBS BOARD

Looking for a quant job, risk, algo trading,...? Browse jobs here...