Communicating with a PHP script (and pretending I'm a browser)

Chris Angelico rosuav at gmail.com
Tue Nov 11 16:56:56 EST 2014


On Wed, Nov 12, 2014 at 4:55 AM, Ethan Furman <ethan at stoneleaf.us> wrote:
> On 11/11/2014 09:30 AM, Larry Martell wrote:
>>
>>
>> They are technically savvy. They are a 100% PHP shop. They have a big,
>> complicated app that they've been working on for 10 years. No one
>> there knows python or django. They want to put some new frontends on
>> their app. I was bought in for another project (involving Google Tag
>> Manager and Google Analytics), which I completed. Then they asked me
>> about this project. I told them they should redo their app in Flask or
>> Django. It took some cajoling, but they eventually said OK. But then a
>> few days later they said before I went off and reimplemented
>> everything in python, could I just build the new frontend and call the
>> existing PHP code. This would enable them to get the new frontends out
>> to their clients sooner, and then I could go back and port the PHP to
>> python. I don't see what is so wrong with that.
>
>
> Sounds like an excellent game plan to me.  :)

I wouldn't go so far as "excellent", but certainly it's not
nonsensical. It's just a question of balancing the cost of the
hybridization (development effort that, once the Python rewrite is
complete, will be discarded) against the time gain of getting
*something* sooner. Also, this will probably have a significant impact
on performance/throughput, so everyone needs to be aware that
benchmarking Python by looking at the new app's performance would be
horribly unfair. But this is a viable proposal.

ChrisA



More information about the Python-list mailing list