There seem to be a lot of questions in the thread regarding how this thing works, one commentator noted that you could just bind something to window.onerror (which is a valid point)
They basically seem to have their own JS asynchronous loader script that will wrap all of your scripts into its own thing. That also means you can't use a JS loader like head.js, require.js etc, but their loader serves a similar purpose. (If I'm not misunderstanding)
I have no idea why they need to do this, they might be extending JS function prototypes to log function calls etc to get a stack trace. Who knows. Way beyond myself atm.
This is just got released, but still there are a few things that strike me as odd/worrying:
- If they really need to load the JS for you, what happens to inline JS and JS that needs to be at the end of the document (for some reason or another)..
- Their JS file is +100k un gzipped.. That seems a lot if you compare to any other JS loader (typically in the 5-10 k range
- Note: 100k of JS to load in the head is a LOT of KB before the user starts downloading some HTML.
- They are actually including all of jquery within their JS file - not in global scope - which seems really wasteful, because you will probably also be doing that - there might be more libraries in there, I'm not sure.
- This taken from their docs:
> 2. We recommend that you download proxino.js and host it locally. This will ensure that your javascript always loads, even if our proxy goes down (which has never happened). Here is the basic form:
"Which has never happened" -- didn't they JUST launch? That seems a bit immature.
I mean, this is probably just their initial beta thing and I'm guessing this will be pretty decent in a year or two, but so far it's just an interesting tech demo, imo.
There seem to be a lot of questions in the thread regarding how this thing works, one commentator noted that you could just bind something to window.onerror (which is a valid point)
They basically seem to have their own JS asynchronous loader script that will wrap all of your scripts into its own thing. That also means you can't use a JS loader like head.js, require.js etc, but their loader serves a similar purpose. (If I'm not misunderstanding)
I have no idea why they need to do this, they might be extending JS function prototypes to log function calls etc to get a stack trace. Who knows. Way beyond myself atm.
This is just got released, but still there are a few things that strike me as odd/worrying:
- If they really need to load the JS for you, what happens to inline JS and JS that needs to be at the end of the document (for some reason or another)..
- Their JS file is +100k un gzipped.. That seems a lot if you compare to any other JS loader (typically in the 5-10 k range
- Note: 100k of JS to load in the head is a LOT of KB before the user starts downloading some HTML.
- They are actually including all of jquery within their JS file - not in global scope - which seems really wasteful, because you will probably also be doing that - there might be more libraries in there, I'm not sure.
- This taken from their docs:
> 2. We recommend that you download proxino.js and host it locally. This will ensure that your javascript always loads, even if our proxy goes down (which has never happened). Here is the basic form:
"Which has never happened" -- didn't they JUST launch? That seems a bit immature.
I mean, this is probably just their initial beta thing and I'm guessing this will be pretty decent in a year or two, but so far it's just an interesting tech demo, imo.