Hi all,
Hope I can form this into an appropriate question, but basically I am looking for answers to the question:
Do you find the ability to receive an e-mail of results from a webserver useful?
An ancillary question is do you selectively not use webservers (when open alternatives are available) that require you to submit an e-mail address to run the results?
We are currently developing a new webserver and the results take in the order of 5-10 minutes to calculate for a given protein sequence. It does not currently have any kind of REST or SOAP type interface (but will have downloadable version for scripting).
Thanks for your time.
I believe that for "complex operations where the time cannot be reduced below minutes" the webserver is actually a perfect solution. Take Haddock docking program. Using the program locally requires lots of software to install, and then enough computational power to get the results. But they provided a web server (http://haddock.science.uu.nl/services/HADDOCK/haddock.php) that has excellent easy to use interface, and distributes your jobs over a grid. Much better solution that local installation.
Yes, but how many web servers have a grid or HPC solution at the backend? Answer: very few. Most bioinformatics web servers are single machines with no consideration of scalability.