SERVING THE QUANTITATIVE FINANCE COMMUNITY

  • 1
  • 3
  • 4
  • 5
  • 6
  • 7
  • 49
 
User avatar
Cuchulainn
Topic Author
Posts: 59715
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Three years from now: can you predict the Software Landscape in anno 2018?

September 5th, 2005, 2:57 pm

QuoteOriginally posted by: NOf course, mgt. decision without involvement of IT staff is also not good.Software engineers don't like being called 'IT staff'.You should strive for Nash equilibrium: good for me AND good for the group.My IT staff calls this 'win-win'.NPlease give my apologies to your team. I have been working since 1978 with computers and would you believe it, I still do not know what SE is? Just imagine that.I hate buzz words. Some more are:mindsetin the loopworst case scenarioPrioritizewin-win (usually means someone wants to borrow something)etc.
Last edited by Cuchulainn on September 4th, 2005, 10:00 pm, edited 1 time in total.
 
User avatar
ZmeiGorynych
Posts: 876
Joined: July 10th, 2005, 11:46 am

Three years from now: can you predict the Software Landscape in anno 2018?

September 5th, 2005, 3:02 pm

QuoteOriginally posted by: CuchulainnYou should strive for Nash equilibrium: good for me AND good for the group.Now do I remember correctly that the Nash equilibrium is when it is not rational for any one party to deviate? Thus in the Prisoner's Dilemma, the Nash equilibrium is when both prisoners confess - not exactly a win-win situation, more like "If I'm suffering, so will you!"Still a workable metaphor for management-engineer relations perhaps?
 
User avatar
alohashirt
Posts: 60
Joined: July 14th, 2002, 3:00 am

Three years from now: can you predict the Software Landscape in anno 2018?

September 5th, 2005, 3:09 pm

QuoteOriginally posted by: lballabioAs for dynamic languages becoming mainstream: this page lists the number of projects per language at the Sourceforge site.I openly admit that it's a biased sample---they're all open-source projects, so they don't give any hints about industry penetration---and that it's of course impossible to detect any trends with a single point in time. But it might be interesting to see what languages programmers use when they are left free to choose their tools. They should be the experts, after all...LuigiWow I am really surprised by the number of Perl and Python projects. I didn't realize how much penetration Python already has with developers. I am bullish on dynamic languages for a couple of reasons: verbositymy prejudice is for static typed languages a la Java C# but I have seen the "The Law of Unintended Consequences" kick many many organizations in the butt over the last five years. Java is a much easier language for an average developer to use than C++ and the result has been a diarrhoea of source code filling up CVS repositories across organizations. If we created some notional signal to noise metric for this code and measured it we'd see that in typical enterpise applications less than 5% of the code is business logic. The bulk is the dross of marshalling data to and from JDBC result sets, plumbing, configuration, resource management - cruft and make-work. I have just started work on a trading system that supports *two* fixed income products. Based on my understanding of the function ality I'd figured it would be about 30k loc. It is 800kloc like some crappy Cornish Pastie that is all pastry and only a smell of meat. I'm not naive enough to suggest that a weakly typed language would fix this. But it can help.ExperienceI have seen one broker dealer do extremely well with derivative pricing and risk apps written in a dynamic scripting language - they have a real competitive advantage through responding to new business quickly. I've also noticed a bunch of smart programmers / authors abandon Java for Python, Ruby, Groovy, some other in the past four years.
 
User avatar
exneratunrisk
Posts: 3559
Joined: April 20th, 2004, 12:25 pm

Three years from now: can you predict the Software Landscape in anno 2018?

September 5th, 2005, 3:17 pm

provocatively speaking, I think projects have a good chance to be effective and run efficiently if the approach is uncompromisingly: given a problem, searched for its adequate decomposition, methods and tools . Not the other way around (what is the use of functional programming if the problem is clearly data-oriented?)This is why I, as manager, "prefer" senior programmers having "multilingual" programming skills. Java, C#, C++, Mathematica, scripting languages,....In another thread here the question arose "VBA in .NET?". I, interested myself, asked my senior developer. He does not know VBA well and he is not a .NET expert either. His immediate answer: "why does one want to use a simple language in a complex environment? In combination they both lose their advantages."Our real pain: coping with all type of operating systems in all runnig versions. Win, OSX, Linux, Solaris,....(BTW, I also like developers, who are able to create task-oriented, special-purpose programming languages. Why? Most of the complex tasks are not of the point&click nature).
Last edited by exneratunrisk on September 4th, 2005, 10:00 pm, edited 1 time in total.
 
User avatar
Cuchulainn
Topic Author
Posts: 59715
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Three years from now: can you predict the Software Landscape in anno 2018?

September 5th, 2005, 3:45 pm

QuoteOriginally posted by: ZmeiGorynychQuoteOriginally posted by: CuchulainnYou should strive for Nash equilibrium: good for me AND good for the group.Now do I remember correctly that the Nash equilibrium is when it is not rational for any one party to deviate? Thus in the Prisoner's Dilemma, the Nash equilibrium is when both prisoners confess - not exactly a win-win situation, more like "If I'm suffering, so will you!"Still a workable metaphor for management-engineer relations perhaps?yes, u are right. If you mess around, you get punished on the next game. Say it to my kids all the time. They know that zero-sum games do not exist.Win-win does not exist, at least not in my neck of the woods. They are all out to get me> when both prisoners confess - not exactly a win-win situation, more like "If I'm suffering, so will you!"Another good option is for the prisoners to keep their mouths shut. It is sometimes better not to cooperate at all. The fallout is sometimes less. That's also in the Nash model.> Still a workable metaphor for management-engineer relations perhaps?Each party should know what the payoff is in each scenario. Acrimony or greed will come back to haunt you. BTW you can program it in C++ , use LCP (quadratic programming), same technique as for early exercise (see my book page something). Then plug in your prisoner data
Last edited by Cuchulainn on September 4th, 2005, 10:00 pm, edited 1 time in total.
 
User avatar
Cuchulainn
Topic Author
Posts: 59715
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Three years from now: can you predict the Software Landscape in anno 2018?

September 5th, 2005, 3:49 pm

> It is 800kloc like some crappy Cornish Pastie that is all pastry and only a smell of meatI like that one
 
User avatar
DrBen
Posts: 499
Joined: February 8th, 2003, 1:24 pm

Three years from now: can you predict the Software Landscape in anno 2018?

September 7th, 2005, 8:32 am

alohashirt: Great post and I go along with you 99% of what you said,however you mentioned:Quote...products of the OMG...now as far as I know OMG do not produce products but only specificationsfor which they may or may not produce a reference implementation. Moreover,I would say that the specifications they have produced has been the formulationof ideas which have already existed within products and programming practicesfor many years before in similar guises.Quote...Corba, UML, XMI, MOF...Are all great "ideas" in the broad sense, however the resulting products so tospeak often where not suitable to be used by humans. For example, Rational Rosewas a huge bloated whale, which took forever to install, was slow, took up 4GB,and all I got it to product was pictures I could have produced with MS Paint.The Corba spec is 10,000 pages (need I say more) but the (partial) implementationsof that spec certainly could do everything that Web services do now 10-20 yearsago.QuoteN: Who cares. In ten years, I'm sure all software will be outsourced to India.I hope not but it is a worry. My worry lies in the fact that even if this doesnot happen and the salary levels just relatively devalue them a reasonable portionof the talent is just going to going into other industries.Anyway, at the end of the day it is products which will actually change IT andto this end technology in products in 10 years are at least ideas in peoples headsnow. Towards this end, and just to get specific (rather than giving a plug) wouldyou guys consider the following talk to contain ideas which will be in IT in 10 years:Dr. Richard Nicholson, CTO, ParemusComplex Adaptive Systems, Jini, andthe Enterprise DataCenter - Friday September 30 2005http://groups.yahoo.com/group/jsig
 
User avatar
Cuchulainn
Topic Author
Posts: 59715
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Three years from now: can you predict the Software Landscape in anno 2018?

September 7th, 2005, 9:06 am

Dr Ben,I spent time trying to fathom MOF (Meta Object Facility). To make a long story short and not sound too negative, I doubt if it will be used in the mainstream in the forseeable future. Of course I could be wrong.
 
User avatar
DrBen
Posts: 499
Joined: February 8th, 2003, 1:24 pm

Three years from now: can you predict the Software Landscape in anno 2018?

September 7th, 2005, 10:51 am

QuoteOriginally posted by: CuchulainnDr Ben,I spent time trying to fathom MOF (Meta Object Facility). To make a long story short and not sound too negative, I doubt if it will be used in the mainstream in the forseeable future. Of course I could be wrong.To be honest I never really looked into MOF (I should not have quoted it). I've had much less timesince the IT pick-up over the pass year, hence less Wilmott and less reading. Also, XMI is reallyjust a human readable desciption of UML diagrams as opposed to gif which is a machine onlyreadable form, hence not really a "great idea" but certainly a useful addition.
 
User avatar
Cuchulainn
Topic Author
Posts: 59715
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Three years from now: can you predict the Software Landscape in anno 2018?

September 7th, 2005, 11:00 am

QuoteOriginally posted by: DrBenQuoteOriginally posted by: CuchulainnDr Ben,I spent time trying to fathom MOF (Meta Object Facility). To make a long story short and not sound too negative, I doubt if it will be used in the mainstream in the forseeable future. Of course I could be wrong.To be honest I never really looked into MOF (I should not have quoted it). I've had much less timesince the IT pick-up over the pass year, hence less Wilmott and less reading. Also, XMI is reallyjust a human readable desciption of UML diagrams as opposed to gif which is a machine onlyreadable form, hence not really a "great idea" but certainly a useful addition.I use a subset of UML before We start designing and implementing: Class diagrams Component diagrams Sequence diagramsThese are v useful and reprsenet 30% of UML 2.0. The other 70% has less added value. Maybe useful for large teams, (50 and more)
 
User avatar
TraderJoe
Posts: 11048
Joined: February 1st, 2005, 11:21 pm

Three years from now: can you predict the Software Landscape in anno 2018?

September 11th, 2005, 7:41 pm

QuoteOriginally posted by: CuchulainnQuoteOriginally posted by: NOf course, mgt. decision without involvement of IT staff is also not good.Software engineers don't like being called 'IT staff'.You should strive for Nash equilibrium: good for me AND good for the group.My IT staff calls this 'win-win'.NPlease give my apologies to your team. I have been working since 1978 with computers and would you believe it, I still do not know what SE is? Just imagine that.I hate buzz words. Some more are:mindsetin the loopworst case scenarioPrioritizewin-win (usually means someone wants to borrow something)etc.You'll never make a good manager .
 
User avatar
Cuchulainn
Topic Author
Posts: 59715
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Three years from now: can you predict the Software Landscape in anno 2018?

September 12th, 2005, 5:22 am

> You'll never make a good manager .My IT staff are self-managing.I never procrastinate until tomorrow anything that I can delegate to someone else today.
Last edited by Cuchulainn on September 11th, 2005, 10:00 pm, edited 1 time in total.
 
User avatar
abginfl
Posts: 38
Joined: February 2nd, 2005, 2:03 am

Three years from now: can you predict the Software Landscape in anno 2018?

September 13th, 2007, 4:43 am

It's been just over two years since this thread lost its momentum, and I was wondering if people still hold the same opinions on the nature of future software development given changes in hardware that are here now or coming (multicore, gpu's, fpga's, special purpose processors-i.e. Cell BE, accelerator boards, manycore processors, etc.)? I was reading an interesting blog on the topic here. So how much do you all think the changes in hardware will effect how everyone writes code (from QF/HPC/C++ gurus to back-office code monkeys)?
 
User avatar
Cuchulainn
Topic Author
Posts: 59715
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Three years from now: can you predict the Software Landscape in anno 2018?

September 13th, 2007, 9:22 am

QuoteOriginally posted by: abginflIt's been just over two years since this thread lost its momentum, and I was wondering if people still hold the same opinions on the nature of future software development given changes in hardware that are here now or coming (multicore, gpu's, fpga's, special purpose processors-i.e. Cell BE, accelerator boards, manycore processors, etc.)? I was reading an interesting blog on the topic here. So how much do you all think the changes in hardware will effect how everyone writes code (from QF/HPC/C++ gurus to back-office code monkeys)?My own subjective and objective list of will happen and should happen (they are randomly sorted is:1. multicore will be standard in h/w but how do we use it? (like when we went from DOS to GUI, anyone remember?)2. parallel and multi-threaded processing will enter main stream, slowly3. C++ evolves and is the de-facto standard here, C# is the rising star4. Software design will become more important5. bigger computers means we will be able to model bigger problems6. OO needs to be subsumed in new ways of designing systems: OO is too fine-grained for high-level initial design7. Software development becomes multi-paradigm: objects, generics, components.Whatever happens, I think that developements in computing power will influence many of the above points.Disclaimer: these are my personal views; prediction is very difficult, especially predicting the future
Last edited by Cuchulainn on September 12th, 2007, 10:00 pm, edited 1 time in total.
 
User avatar
zeta
Posts: 1952
Joined: September 27th, 2005, 3:25 pm

Three years from now: can you predict the Software Landscape in anno 2018?

September 13th, 2007, 11:37 am

I hear the following are destined for a comeback:a) vacuum tubesb) punch cardsc) algold) 2400 bpse) wire wrap
ABOUT WILMOTT

PW by JB

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


Twitter LinkedIn Instagram

JOBS BOARD

JOBS BOARD

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


GZIP: On