Quantrix Feedback – Need for Speed
I’ve been using Quantrix now for 2 years. When I did my upfront research I was looking for a tool to become a replacement for Excel for the next 20 years. I felt that the Excel had matured, and was hampered by the limitations of the spreadsheet structure.
I’ve built models to assess large portfolios of moviecontracts in depth, to calculate large telecom fiber rollouts and capes, I’ve done montecarlo simulations on options and some regular base projection financial modelling.
Since my start I’ve seen Quantrix evolve with DataNav and DataPush to become a much more powerful platform, very much a direction in which I hoped Quantrix to evolve.
I have three wishes for future development and I’d love to hear other people’s view on this.
1. Speed.
First an example: I envision that at some time in the future we’re able to run Monte Carlo simulations on business plans on the fly – at that point we’re not talking just scenarios but truly capturing the flux in external developments and showing financial projections within statistical ranges. With today’s Quantrix this is not feasible. The building blocks in terms of structure from a modelling perspective are there, but the speed isn’t to make this practically feasible. As improvements in computing speed these days are coming from increases in cores, and computing-intensive Quantrix tasks (in general) are highly parallel (for example: Monte Carlo simulations), Quantrix speed should increase significantly if it could tap multiple cores. I’m talking general parallel processing, and not some special client-specific installations.
2. Make the QAPI setup more accessible.
I’ve looked at the QAPI setup, but I find it to be too intimidating for a business user as myself. I think more people would be able to tap into this powerful tool, if it were more easily accessible. Maybe I should go for a course, but I’m unsure how steep the learning curve is. I’m a business user, and I’d need to be able to overcome the base threshold within a single course.
3. Less restrictions on the viewer
I need to get models out to people who don’t have a full Quantrix installation. In order for them to be comfortable with the models, they need to be able to play around with parameters to verify assumptions and to understand the logic. As far as I can see now, they can’t.
I’d like to suggest they should be allowed to make changes, but not be able to save them – to avoid Quantrix cannibalization. If you need something more restrictive, just let them make changes on the Presentation canvas only.
These are my thoughts. I’m looking forward to the input of other people (and of Quantrix staff if possible).
I agree with all your points! I’ve been looking at 1 and 2 for a while now and on 1 I have some ideas for a compiler (could deliver speed and parallelism but at the expense of user interface) and on 2 I “solved” this by writing a plugin which exposed the QAPI via the jython scripting engine (so new plugins can be written in Python rather than java, thus sidestepping the compilation process).