As new technologies continue to emerge, many are looking for an answer to this question, particularly in strategy and product development departments in brokerages around the world.
Everybody have seen how other retail industries have moved onto web and mobile, and the general consensus seems to be that trading platforms are destined to head this way as well – in fact, it almost seems as if people are taking this for granted.
While I don't necessarily disagree — I'd like to try and challenge that thought for a minute.
I think it's very true that mobile/web is currently the preferred media for many platforms that have a broad mass retail audience and where the tasks carried out are relatively simple (facebook, Spotify, Viber). However that's not necessarily the case when it comes to professional tools (Photoshop, Dreamweaver, MS Office−suite).
These tools are still better suited for the desktop, mainly due the lack of screen real estate on a smartphone or even a tablet device, and because they need access to a lot of computer power to do some very advanced things in the background, which are still not possible (to the same extent) using web technologies.
The questions then become: a) how many advanced things does a Trading Platform have to do, and b) is a trading platform in fact a retail or professional tool? (User experience−wise, forget MiFID categorizations)
As for a, the answer is A LOT. Ironically, in the vast majority of cases, the more simple you make a trading experience, the more code you have to write and hence the more computer power you have to access.
As for b, well, traditionally the trading platform has been a professional tool − no question there. FX and trading in generally is certainly becoming a more retail/mainstream activity, however, out of the $300 billion being categorized as "retail FX volume", I think it will take a while before the vast majority of that comes from a truly mainstream audience of traders.
Finally, there are a series of user−experience issues one should consider just taking for granted that absolutely everything will run on web in the future. Consider why products like Spotify, GoToMeeting, join.me, Skype, Dropbox and basically all software that Apple ever made has been done as thick applications and/or has their thick applications seeing much more adoption than their web counterparts.
HTML5 is a fantastic buzzword (that few people actually know what mean), but the experience that one can provide through HTML5 (unfortunately) still seems to be somewhat inferior to that of native apps (ask Mark Zuckerberg, he'll tell you).
When I say that I don't necessarily disagree, it's because I do see a future where all the tasks we do are increasingly carried out on mobile and web.
The stats don't lie, many products and services are moving to mobile/web. However, I think that specifically for something like trading, it will at least be another couple of years before web and mobile becomes more relevant than desktop applications, and I think its a change that will only happen as retail FX becomes even more mainstream.
In short, the desktop platform will continue to be relevant for the years to come. Of course, this depends on your user/client base – but I'd like to try and see someone try and write an algo on an iPhone.