Box.net CEO: HTML5 could kill desktop software
venturebeat.com
venturebeat.com
- Matlab/Mathematica - Photoshop - Blender - America's Army - OneNote - LinqPad - ...
(bring on the pain)
What would putting it under web page do for it? People might get confused that they can switch machines and continue working with their files, while in fact this is not possible.
HTML5 is cool, but for some (powerful) users plain-old shrink-wrapped software is better, especially if that software is dealing with data close to gigabytes (CAD, Image processing, etc.), or if that data is shared adhoc (there are special versions of Maya where instead of saving/loading to the file system, everything goes through a specialized file I/O - this might come a custom solution from Autodesk, or some other way).
I work in game company, and the only web applications that we use are the ones that make sense to be used from there: - Bugtracking - DevTrack or Bugzilla - Some custom statistics server - Intranet pages - vacation forms, etc.
Everything else is native Win32, or Win64 apps - MFC, wxWidgets, Qt, .NET, Direct3D, whatever there is. I doubt we'll switch this way, as moving some of these apps to the web is simply "NEW" to us, and we still don't trust it enough.
Every app you move from shrink-wrap version to the web - you need a server - every server in an organization needs someone to take care of it - what happens when it needs to be rebooted? what happens if it goes down?
It's way safer for certain apps, to be as they are.
For example I can't wrap my head of using P4 (Perforce) over the web interface - yes it's there - but I prefer (the obsolete now) P4Win32.exe
The real point of porting to the web is to make these apps accessible to more people. To apply web workflows to their development, like change their UI after monitoring analytics, ship often, etc..
Large production apps like Photoshop can be split off into thousands of modules, which will be loaded only when the user is using a particular feature. That's a very different development process, currently Photoshop is developed for CDs, one big ship date, one big upgrade, all features in, want them or not.
The way apps are paid for could also change. If you're not using certain features, you shouldn't be paying for them, shouldn't be investing in their development while features you do use are neglected.
It's not really the technology of the web that will bring about such changes. NaCl mainly enables you to run normal binaries, as you already do on the desktop. The real change will come from people rethinking how they make and deploy apps.
http://www.sagemath.org/ http://www.monkeyanalytics.com/
I am not sure what "HTML 5" has to do with analytics software though. I usually don't need video or crazy mouse capabilities when doing statistical analysis.
I guess what the original poster meant, was instead of plugin use HTML5, but I dunno whether that's going to work for Mathematica (performance wise).
Well, I don't know how much we have to worry about production oriented software disappearing.
From what I see, the web has caused a shift in the production of cultural content, away from centralized media industries to something resembling cottage industry. Wouldn't there be a rise in production oriented software to meet the demands of creators needing better, faster, and easier tools?
Let's face it, the web is never going to take over for production oriented software. People will need to be squeezing every last calculation out of their hardware while they're making things and they need feedback without delay.
You couldn't work in Illustrator with a 50ms delay, so where do you draw the line for what is computed client-side and what on the server? Sure, filters could be processed server-side. But complex drawings, taking up lots of RAM, constantly requiring a re-render?
We don't even need to talk about server-side audio production.
It's not that there aren't tools that work just as well or much better when served in the cloud, it's just that there will always need to be tools that demand to be run as close to home as possible.
It seems to me that the amount of people wanting to use production oriented software is going to grow at least at the pace of consumer oriented software.
My personal view is that the web is good at things which are inherently communication centric. Nothing else justifies the restrictions that working within the web browser brings. Flash might be the whipping boy of internet technologies, but at least its reasonably write once.
I think the thin client architecture is a paradigm that will constantly cycle in and out of fashion but never quite deliver on its promise.
HTML5 will be the death of Flash and Silverlight as we know these products today. That doesn't mean the product by name won't exist (think jQuery-esque like platforms for HTML5), it will just change forms.
That's my guess anyways.
I'd be curious to know how normal this is. Maybe if a lot of websites implement it, we'll be able to get some real stats on how often it actually gets used.