mercredi 24 février 2016

Where to start when planning to develop web applications [on hold]

I'm an old time programmer. Came from Basic (old apples), Pascal, Turbo C, then GCC/C. I started as "web developer" when PHP was first launched as language (PHP3), in that time we have a fight war. Mozila didn't existed. We used Netscape 3 and 4. And IE ruled the market. My life coding to web where a living hell, nothing worked as expected the standards were not followed. From there to now, I have just coded system application, with none or low interactivity mostly to Linux or system libraries. Today my main languages are Bash, C and Python.

But now I'm starting a small project (small in size, but maybe big in intentions). I have started to write the libraries and they are working fine. When I stumbled over the idea to create a GUI, and I was thinking in PyQT, PySide or PyGTK (I like GTK interface), but another possibility came to my mind (why not web based). And hence my question. I had a awful experience long time ago, and today there are so many technologies that I don't have idea where to start.

About my project:

Is a scientific application. I get some bulk data (for example from a sensor map), the scientist render this map on screen (is a image), right now I'm using the amazing matplotlib. Tinker with some parameters (as maxima and minimal values, ranges and such). With that values it will test if the parameters are ok, that means use that parameter to render some "random" data from the whole datasource.

I will display (in the interface) the test data in some modes (side by side, animated, grouped).

If the test was approved. I will render the whole data source based on that parameters.

Right all rendering stuff and such is done, but the magic will be on interface (and the web bring the possibility to do it from anywhere from any device) that will make the scientific analysis faster (because get an directory with 50 images and compare them is the real problem, and if not ok, get back and render all again).

Today I use mostly pandas, numpy, scipy, matplotlib.

I'm looking for ideas and methodologies to evaluate and check how to choose the right technologies that will help me to get faster and better on my objectives.

Aucun commentaire:

Enregistrer un commentaire