For what it’s worth, if they use a source code control system (e.g. git) with some discipline (e.g. useful check in comments) then there are tools which will write release notes automatically for each deployment. Of course, someone should go translate from ‘programmer’ to ‘user’ language, and filter out trivia, but at least you have a comprensive starting point. Given that they’re using Google App Engine, which by default does the right thing, they should have this.
Admittedly web apps tend not to expose version numbers, but for a tool like this I’d think they’d want to capture the version number in all bug reports, etc., which you can do by putting the build number into the app URL. Makes all sorts of troubleshooting easier, too.
We post changes to Latest Improvements. It’s not perfect (there’s always assorted small bug changes and performance improvements - I guess we could just say that every week) but that’s where you’ll find all the notes on our latest release. We also roll non-app changes in there, like the spare parts availability we just announced.
To echo what @johnse is saying. They absolutely are watching their resource contention and sizing as appropriate. I’d even imagine they broke out the different pieces of the solution into smaller services. Renderer, master DB server, application / front-end servers, then some form of load balancer up front. Their renderer / slicer app servers are likely pretty large machines, whereas the UI / DB may be a touch smaller. The most important thing to remember is there isn’t one ‘server’ that is doing all this. A properly architected SaaS solution is going to be in the dozens / hundreds (as needs grow) of servers / vms.
Cloud solutions are fantastic for non organic growth models, and unknown workloads. They can scale up and out as needed and then scale down for savings during off hours and what not. Once they get a sense of what the load is going to be; they may bring it in-house to get some cost savings. (Cloud providers are not cheaper…they are more flexible). But throw the platform on some form of Hyper Converged system in-house and it’d probably knock our socks off with how performant it is. But until the growth models / plans are more predictable…it will be sized in the cloud at a balance of pricing / performance.
They do get it. They always seem to ask for details on exactly when the problem happened. That, coupled with the identity of the user, gives them the version number.
To add onto @palmercr ‘s post, I can almost guarantee every device that anyone has ever had issues uploading a to the glowforge servers with, has the horsepower to do freakin line tracing and camming, locally, if they chose to go that way. If I can do augmented reality in real time at 60fps from my smart phone, it can handle an illustrator function that’s existed for decades.
How many companies make software that works on Windows, Apple computing devices and Linux devices out of the starting gate? I agree local software would eliminate some of this complexity. But the reality is many devices are limited to certain computing systems. And Linux need not apply.
From Epilog:
Operating System: All new Epilog lasers are designed to work with all versions of Windows 2000, XP, Vista, 7,8, and 10 operating systems.
The Fusion Laser Series is also compatible with Mac operating systems OS X 10.7 and higher. Read more about our Fusion Mac Driver here.
And software to use the Silhouette Cameo are only Windows and Mac and some iOs support. No Linux.
At the moment all the things that I can do with the Glowforge in its present state of software development on the four different computing systems I use almost every day outweigh any of the limitations with engraving sizes. These are my needs and engraving just isn’t the biggest use for me.
Would love to have anyone point me to a startup company that provides functional software that runs natively out of the box on all four computing platforms. Five if you throw Android in there.
Granted, it’s not complete, But Chrome on the main three get the job done.
I will point to some game companies that do support them all…but only because the Unity 3D development system supports about 29 platforms
I completely agree with your point. Even with an awesome tool like Unity, it takes work to use each platform effectively. Cross-platform development is hard…but it is getting easier. C#, .Net, along with Xamarin is getting to the point where something like a local GFUI could be done quite effectively with a single code base across those platforms.
I don’t think the point was about actually running it on random devices. The point is in this day an age the devices we use like phones, tablets and even my watch appear to be more powerful than their cloud. It is completely crazy that a cloud server takes 20 minutes to calculate a 2D tool path and runs out of resources if it is too complex. An RPI can slice a 3D part in that sort of time and each layer of many hundreds is about the same complexity as a 2D print. And it does kerf compensation!
