The Price Point of Software?

[NickK]

Software is inherently costly to produce. No ifs, no buts, no free lunches.

The software engineering wisdom is that you generate something that does the core minimal functionality that is dependable then build on it. As software is akin to building a city with each building in intricate detail by a few people - it doesn't bode well if they bite off more than they can chew.

There's a also the perception of value associated with price. So dropping the price does not often equate to a higher volume of sales but instead appears that the value to the user is less and so the user does not buy (because expensive always sets the expectation that it's going to be awesome!). So it may look like there's little in something free and something low price.

There is also the leaving the money on the table argument - if the consumer values something then there's no point in leaving potential revenue on the table. For niche products then the return is normally factored to make the most out of each sale because the market is so small there's no information on just how many sales there could be.

Now this leads to the ultimate showstopper - the price war. This doesn't benefit anyone.. as the number of products fall and exit the market, quality goes to pot, the support is dire and you're then paying for each minor increment.

Open source and support? Well not many people will pay for product support only product.. then mutiply peanuts by peanuts and only the monkeys are happy.

Apple have shown that people will continue to pay highly for products that work, that allow the user to focus on their task (and not spend their time fixing it) - happy to effectively pay for software as a service without it being advertised as such. Now Photoshop made it blatant with their monthly subscription and the users shock caused an outrage.. 

[perfrej]

Hi Per,

had a look in the ansvr code and it looks like it uses the exact same principles with the online API version.

Both need to be accessed by POST and then

-- log in

-- upload image

-- pass parameters wrapped in json format (emulating webform) for solving and waiting for results.

They also have a python client ready here -> http://trac.astrometry.net/browser/trunk/src/astrometry/net/client/client.py which i tested and worked (partially cause i know jack about python atm) and got to the point of uploading an image with specific parameters to the local server.

Which method - way have you tried?

Hi there!

I have checked out the entire astrometry.net code now and have a working VM with all but the web service. I tried ansvr with my own code and with the standard python astrometry.net client. The ansvr server hangs after image upload and then has to be restarted. I do not think it is based on Dustin's original web service code, which by the way runs on a different server than the main solver.

I am a bit short of time now, but my client code runs perfectly fine against the real astrometry.net web service so everything points to ansvr not behaving.

/per

[frugal]

I have always been amazed that astronomers will easily spend £1k on a mount, £2k on a scope, £1.5k on a camera and hundreds more on filters, cables and laptops... but ask them to spend more than 50 quid on software and they balk.

[frugal]

Apple have shown that people will continue to pay highly for products that work, that allow the user to focus on their task (and not spend their time fixing it) - happy to effectively pay for software as a service without it being advertised as such.

Unfortunately, Apple also taught people that 99p is the correct amount of money to pay for software...