Counter to Thomas’ instruction, I attempted many different things: first trying out several different Mashape APIs without success before attempting to fiddle around in Aframe VR (since we had a still-forthcoming homework assignment using this framework), also without much success. For the discarded APIs, in the end, I wasn’t really sure what the APIs did; there wasn’t much documentation to be had (or maybe I didn’t know where to look), and they seemed harder to execute than the Yodaspeak example we did in class. When I did get a server response, it was an error. I think it was a 503 error (Service unavailable); even after looking this up, I still had no idea how to fix it, or if it was even a problem with my work or with the API itself.
So, into the Aframe VR experience I went. Looking through the work we did in class, there wasn’t a whole lot of code to be written, and it seemed fairly user-friendly. There was a good amount of documentation, though most of it not current with version 0.4.0. I tried several tutorials, following them line-for-line while substituting my own images. Only about 50% of the time did all of the assets I brought in to index file appear. I did successfully bring in a few equirectangular images to create 3d images, but couldn’t achieve anything beyond that. And since bringing in the image is literally just one line of code, I couldn’t feel too happy with that. I’ll have to go back to that next week.
And so, back to the APIs I went. I found yet another API to try out (a grammar app) and after much tinkering, I did manage to get a response that wasn’t an error. It wasn’t the response I was expecting to get, so I asked my husband to help. He helped me to get over the hump so that I was able to get it working. I prettified it in class today, and added a very modest CSS animation. View my app here.
In order to improve this app:
- Must-have: currently the user will get the same response whether there is no text entered for grammar checking or if there are no grammatical errors. This is an obvious flaw and some more specific conditions should be put in rectify this.
- Nice to have: sometimes the app doesn’t provide a useful response; it often just says “there may be a problem”. With such a vague response, it would at least be helpful if the app could indicate which words it suspects as being incorrect.
- Nice to have: a better loading animation. I had envisioned something else but it is beyond my skill level.
- Version 2: what would solve many of these issues is if the API were a little better and gave the same amount of detail for each suspected grammatical error.