11–17 of 17 posts
I'm going to geek out here for a minute at the fact that Facebook made a screenshot diffing tool for testing. I've been wanting to build that very thing for a long time, but every developer I've talked to about it looks at me like I'm crazy. So, am I still crazy for thinking screenshot diffing might be an effective way to verify builds?
I'm going to geek out here for a minute at the fact that Facebook made a screenshot diffing tool for testing. I've been wanting to build that very thing for a long time, but every developer I've talked to about it looks at me like I'm crazy. So, am I still crazy for thinking screenshot diffing might be an effective way to verify builds?
HTML5 is not effective at the moment, but as phones become more powerful. I think it will take over. The same way the web app took over from desktop apps when personal PCs and network bandwidth improved.
Perhaps, but the same has been said for what the last two or three years and in that time frame we've gone from single core to quad core and 800 MHz to like 1.9 GHz and HTML5 apps still are not up to snuff. I wanted HTML5 apps to be awesome because the developer experience was easier, but it makes for a sucky product, so I started just digging into native and making better software. I've learned my lesson. At some po…
Seems like no one's really struggling to scale mobile I suspect because 1) usage, although large, is less intense and 2) much less goes over the wire (usually just data, not pages).
I think mobile app development has lots of similarity with 'boxed' products. Microsoft invests a huge amount of money in its test infrastructure for its boxed products, because once it's shipped it's very difficult to fix. I think the mobile development community incorrectly tends to treat development more like the web than they do a boxed product. For reference, more than 15% of Facebook users are using a release from 2 years ago (I will try to find the source for this info).
Earlier quoted context omitted.
Perhaps, but the same has been said for what the last two or three years and in that time frame we've gone from single core to quad core and 800 MHz to like 1.9 GHz and HTML5 apps still are not up to snuff. I wanted HTML5 apps to be awesome because the developer experience was easier, but it makes for a sucky product, so I started just digging into native and making better software. I've learned my lesson. At some po…
With the benefit of hindsight, do you still think the developer experience is easier with HTML5, or was it just that you were already familiar with HTM5, but weren't familiar with the native tools?
HTML5 is not effective at the moment, but as phones become more powerful. I think it will take over. The same way the web app took over from desktop apps when personal PCs and network bandwidth improved.
> The same way the web app took over from desktop apps You just made that up , it's not backed by any fact.
I'm going to geek out here for a minute at the fact that Facebook made a screenshot diffing tool for testing. I've been wanting to build that very thing for a long time, but every developer I've talked to about it looks at me like I'm crazy. So, am I still crazy for thinking screenshot diffing might be an effective way to verify builds?
Originally developed at MIT, now freely available and maintained as open-source lib. It offers advanced (partial) matching at the UI level.