Serving the Quantitative Finance Community

 
User avatar
Cuchulainn
Topic Author
Posts: 64316
Joined: July 16th, 2004, 7:38 am
Location: Drosophila melanogaster
Contact:

Does Open Source Software have a future (of kinds)?

February 1st, 2018, 8:56 pm

It's free, run by volunteers who like to do cool things. It's a bit like small shoe factories in the old days

But what  about the important things? For example, when will C++ support modules, how is QL doing etc. etc..

Will open source be assimilated by the big players?  I can't help going back to 1995 when there were 200 software companies doing AutoCAD; few years later there were 5!

Market-driven software is more focused.

What's the future? Will AI take over?

What's the capital of Assyria?
"Compatibility means deliberately repeating other people's mistakes."
David Wheeler

http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
Alan
Posts: 10569
Joined: December 19th, 2001, 4:01 am
Location: California
Contact:

Re: Does Open Source Software have a future (of kinds)?

February 1st, 2018, 9:25 pm

In the last couple years, here's some wonderful open source software I have needed and installed.

1. My ancient web site (optioncity.net) was built in the (now-deprecated) Microsoft Front Page, and I needed a new site for my book. At first I thought I needed to learn Adobe Dreamweaver to construct a proper site. I finally ran across (open source) WordPress: problem solved! (The lead creators and other project coders apparently make their money from backups, add-ons, etc although most of that stuff can be done freely also).

2.  I occasionally need very limited image manipulation. I never need enough to make it worthwhile to buy (well, now rent) Adobe Photoshop, although many years ago I purchased Corel something for that. Anyway, finally ran across (open source) GIMP, which solved my small problems.

3. I like to store multiple backups of thing, so decided I would do local backups of my WordPress site. Needed an FTP client: discovered FileZilla (works great).

4. My second book was written in LaTex, using a commercial integrated editor called PcTex. Only cost $120 or something, but is mostly defunct now. Well, having bought a new computer, I needed a new LaTex editor; discovered (open source) Texstudio  -- wonderful. (And, of course, it is built upon open source LaTex distributions like MikTex).

As you can tell, I am a big fan of open source when it meets the need. Sometimes commercial things are the best solutions for my needs (Mathematica, MS Office, Adobe Acrobat) and sometimes not.  I think this kind of healthy mix of open source/commercial will persist a long time. 
 
User avatar
Traden4Alpha
Posts: 23951
Joined: September 20th, 2002, 8:30 pm

Re: Does Open Source Software have a future (of kinds)?

February 2nd, 2018, 12:24 am

Alan offers excellent examples of the many competent applications available via open source. There's a number of others such as Linux (operating system), Apache (web server), and even an open source variants of Microsoft Office (https://www.cio.com/article/2392232/mic ... ffice.html).

If there's a common application or task that many programmers need to do, chances are that some programmers have created an open source version of it. There's even a bunch of math technical applications (https://en.wikipedia.org/wiki/List_of_o ... athematics) of which I've only used R (which is phenomenal in it's functionality but not very friendly in it's UI).

Market-driven software may be more focused, but it's more expensive and may have more features than a casual user needs.
 
User avatar
lloydsatchwell
Posts: 35
Joined: April 27th, 2011, 4:00 pm

Re: Does Open Source Software have a future (of kinds)?

February 2nd, 2018, 9:40 am

Of kinds. I believe this could be where ICOs/tokenisation might help.

(Note: I am not a crypto evangelist and these are musings rather than conclusions...)

I'll assume we are talking about Free and Open Source vs Market-driven and Closed Source - but appreciate the real world isn't always so binary.

So one paradigm is "Market-driven" software which as you described is focused, and great at solving visible business problems. This model works through recognising the private ownership of the software, capital is raised by the company and its founders embark on turning their ideas into reality. Developers are hired, less disagreement, clearer direction, ideas are acted on, and can be gloriously profitable (or so the shareholders hope). The wonderful capitalist system at work.

The other paradigm of “free & open source” may be driven by figures like Richard Stallman or Tim Berners Lee, where the incentive may be the goodness of their heart, simply because they want to, or some other incentive. GPS and the internet itself came from a strategic military motivation rather than a desire from the broader market. Not necessarily common to work on such projects since they are not driven by profit in a wider context, and at the beginning it can be difficult to see their ultimate utility to society. These open projects exist however and the software industry is rife with similar collaborative, free examples, TCP/IP, even Wikipedia & stackoverflow etc..

Anyone who has been on this planet for more than five minutes will know appealing to a person’s wallet is far more effective than appealing to their heart so perhaps this is where network tokens come in, or where a market can be incorporated with open projects. Developers and crucially users of a technology can earn tokens for participating –bitcoin, ethereum et al.

Maybe ICOs and tokens/coins shouldn't be looking to replace IPOs and Equity, but to incentivise the future of open, collaborative projects. Maybe.
 
 
User avatar
Cuchulainn
Topic Author
Posts: 64316
Joined: July 16th, 2004, 7:38 am
Location: Drosophila melanogaster
Contact:

Re: Does Open Source Software have a future (of kinds)?

February 2nd, 2018, 3:00 pm

OK, Let's take a concrete example: you want to create a large, maintainable (finance) library that can also evolve as new requirements emerge. Which language would you use?

1. Java
2. C++ 98
3. C++11
4. C#
5, Python
6. None of 1-5.

It's allowed to say why :)
"Compatibility means deliberately repeating other people's mistakes."
David Wheeler

http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
Traden4Alpha
Posts: 23951
Joined: September 20th, 2002, 8:30 pm

Re: Does Open Source Software have a future (of kinds)?

February 2nd, 2018, 4:05 pm

OK, Let's take a concrete example: you want to create a large, maintainable (finance) library that can also evolve as new requirements emerge. Which language would you use?

1. Java
2. C++ 98
3. C++11
4. C#
5, Python
6. None of 1-5.

It's allowed to say why :)
The question seems under-defined:

1. Is the library so amazingly complete and powerful that 99% of customers never need to write any code at all (and thus the language of the library can be whatever the developers like most). Do most customers care what language MATLAB or Mathematica is written in?
2. If customers do write code integrating with library code, then what language do these customers use now (and in the future)?
3. What are the kinds of "new requirements" that could occur (e.g., Yet more variety in the pre-existing module categories? Greater combinatoric abstraction? Must run natively on Android smartphones? Must play nice on GPUs?)
 
User avatar
Cuchulainn
Topic Author
Posts: 64316
Joined: July 16th, 2004, 7:38 am
Location: Drosophila melanogaster
Contact:

Re: Does Open Source Software have a future (of kinds)?

February 2nd, 2018, 4:28 pm

Indeed. 
Users don't care as long as their requirement are satisfied, but which languages 'best' realise these (TBD) requirements?
"Compatibility means deliberately repeating other people's mistakes."
David Wheeler

http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
Traden4Alpha
Posts: 23951
Joined: September 20th, 2002, 8:30 pm

Re: Does Open Source Software have a future (of kinds)?

February 2nd, 2018, 4:59 pm

Indeed. 
Users don't care as long as their requirement are satisfied, but which languages 'best' realise these (TBD) requirements?
How about:

If Android, then Java else not(Java)
If Windows, then C# else not(C#)
If millennials, then Python

Note: Windows on AWS is double the price of Linux on AWS so if elastic cloud computing is a big use case, it's better to avoid Windows.
 
User avatar
Cuchulainn
Topic Author
Posts: 64316
Joined: July 16th, 2004, 7:38 am
Location: Drosophila melanogaster
Contact:

Re: Does Open Source Software have a future (of kinds)?

February 2nd, 2018, 5:23 pm

Indeed. 
Users don't care as long as their requirement are satisfied, but which languages 'best' realise these (TBD) requirements?
How about:

If Android, then Java else not(Java)
If Windows, then C# else not(C#)
If millennials, then Python

Note: Windows on AWS is double the price of Linux on AWS so if elastic cloud computing is a big use case, it's better to avoid Windows.
So, OS is the driver of the choice? That's not a good business solution. The requirements are driven by functionality from the business.
"Compatibility means deliberately repeating other people's mistakes."
David Wheeler

http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
Traden4Alpha
Posts: 23951
Joined: September 20th, 2002, 8:30 pm

Re: Does Open Source Software have a future (of kinds)?

February 2nd, 2018, 5:59 pm

Indeed. 
Users don't care as long as their requirement are satisfied, but which languages 'best' realise these (TBD) requirements?
How about:

If Android, then Java else not(Java)
If Windows, then C# else not(C#)
If millennials, then Python

Note: Windows on AWS is double the price of Linux on AWS so if elastic cloud computing is a big use case, it's better to avoid Windows.
So, OS is the driver of the choice? That's not a good business solution.
Maybe, maybe not.

If the "library" is just a loose collection of code like the CD-ROM in the back of "Numerical Methods" then you are right that the OS is irrelevant.

But if the "library" is a complex application for computing/analyzing financial systems with a beautiful UI and extensive integration with the host machine's software & hardware resources, then picking the first OS is crucial.

And if the library is meant to run in the cloud with some web/internet/command-line interface, then maybe the cheapest option (e.g., Linux) is best because it will save high-end users thousands maybe even hundreds-of-thouands of dollars are year.

But you've never said who the customers or use cases are. Are they PhDs in a hedge fund, newbie code monkeys in a bank's IT department, C-suite suits with a dashboard, or iPhone users trading bitcoins? And what are the use cases: R&D, operational analytics, or production code handling all the bank's transactions?

Or is the library going to be jack-of-all-trades-master-of-none?
 
User avatar
Cuchulainn
Topic Author
Posts: 64316
Joined: July 16th, 2004, 7:38 am
Location: Drosophila melanogaster
Contact:

Re: Does Open Source Software have a future (of kinds)?

February 2nd, 2018, 6:28 pm

But you've never said who the customers or use cases are.

Take existing libraries like Opengamma or Quantlib.

Try these features for starters

http://strata.opengamma.io/introduction/
"Compatibility means deliberately repeating other people's mistakes."
David Wheeler

http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
Traden4Alpha
Posts: 23951
Joined: September 20th, 2002, 8:30 pm

Re: Does Open Source Software have a future (of kinds)?

February 2nd, 2018, 7:19 pm

Strata certainly looks interesting but it raises a yuge business decision for the product design.

Strata claims "it does not impose any database, server or middleware requirements." That can be a huge plus if the customer already has these covered. Or it can be a huge minus if it means that the customer has to go figure all that database, server and middleware stuff themselves.

So what do the your customers want: chunks of code they can plug into what they already have or an off-the-shelf , end-to-end system that does everything?
 
User avatar
Cuchulainn
Topic Author
Posts: 64316
Joined: July 16th, 2004, 7:38 am
Location: Drosophila melanogaster
Contact:

Re: Does Open Source Software have a future (of kinds)?

February 3rd, 2018, 12:50 pm

Strata claims "it does not impose any database, server or middleware requirements." That can be a huge plus if the customer already has these covered. Or it can be a huge minus if it means that the customer has to go figure all that database, server and middleware stuff themselves.

I don't think this is a major issue, precisely because we are talking about databases. Client software accesses databases via virtual machines such as ODBC, ADO.NET and so on. Java and C# are strong here.

Open source db virtual machines have less clout IMO because it is in general driven by volunteers. 
Does "native" C++ support db access?
"Compatibility means deliberately repeating other people's mistakes."
David Wheeler

http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
Cuchulainn
Topic Author
Posts: 64316
Joined: July 16th, 2004, 7:38 am
Location: Drosophila melanogaster
Contact:

Re: Does Open Source Software have a future (of kinds)?

February 3rd, 2018, 12:52 pm

yuge business decision for the product design.
It always is.

BTW, I think the choice of OS is not critical.

You anti-MS guys ;) seem to think VS only works under Windows.

https://code.visualstudio.com/download
"Compatibility means deliberately repeating other people's mistakes."
David Wheeler

http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
Traden4Alpha
Posts: 23951
Joined: September 20th, 2002, 8:30 pm

Re: Does Open Source Software have a future (of kinds)?

February 3rd, 2018, 2:19 pm

yuge business decision for the product design.
It always is.

BTW, I think the choice of OS is not critical.

You anti-MS guys ;) seem to think VS only works under Windows.

https://code.visualstudio.com/download
So then probably the cheapest OS for heavy computing is the way to go and that's probably Linux.