I prefer thick client native apps, but experience in the core enterprise space tells me: Enterprise is about making money and keeping the works as least technically challenging as possible to be able to sustain it at lowest input costs. For thin client enabled application distribution is about both technical capability: scaling servers, flexibility, maintenance etc and also reasons Ben states below. Client-Server architectures have evolved along with progressive improvement in network bandwidth to move data / computation / business logic to the server end. As browsers are improving to support more of computation / presentation logic, the pick up a bit in use of HTML5, the argument for thick client / native applications is narrowing. Cloud computing, Big Data processing, Hadoop map-reduce et als these inherently are shifting the computing paradigm more to the back-end reducing the client to just a view screen. That said, thick native client for mostly standalone apps connected / disconnected have picked up with the iStore / Android MP and in future we could evolve a more suitable generic client that can enable intense computation on client / or run off like a thin client - server as required transparently. But can we throw in as many resources to compete with a Google / MS / Firefox in their run up to beefing up browsers as general purpose clients ( thin / thick as you wish it to be ) I would love to see a Pharo Marketplace with native clients running off on Pharo..!.. On Wed, Apr 17, 2013 at 9:00 AM, Ben Coman <btc@openinworld.com> wrote:
Igor Stasenko wrote:
<ranting> i wonder why people so obsessed with web? it is ugly, and dead before born.. </ranting>
Reasonable rant. I'm not a web developer so I am only guessing...
Consumer perspective * Zero Install - or at least the perception of Zero Install. Of course you actually need to install a browser of choice, but commonly these are pre-installed for Mum&Dad. * Sandboxing - executing code with limited access to the system - ignoring security exploits.
Developer perspective * Fast Centralized Deployment - updating the server updates all users. * Common Platform - theoretically reduced effort although issues with compatibility and least-common-denominator
Evil Corporate/Government Perspective * Common Platform - to track _everyone_ and targeted advertising.
cheers -ben
On 17 April 2013 02:19, Ben Coman <btc@openinworld.com> wrote:
Torsten Bergmann wrote:
Thanks - Pharo 2.0 and Zinc makes it really easy to work with, write AND test HTTP stuff:
That sparks an idea, perhaps for a student project... That a general HTTP testing framework for use against non-Pharo systems could be just the thing to help the uptake of Pharo in "enterprises". The level of trust required for an internal testing tool is less than that required for an end-user system. This might be a good stepping stone for enterprises which may initially view Smalltalk a bit strange. Unless of course best practice is to use web testing tools from within real web browsers (I not familiar with what the options are). Perhaps a website stress test tool where the live exceptions can be inspected is something that other tools cannot do.
cheers -ben