With the latest update Firefox has finally dropped the majority of plugin protocols; this includes Microsoft Silverlight. Does FIRST plan on moving from Silverlight to Javascript or a more modern framework? Internet Explorer has many issues from security flaws to crashes and even the inability to use it on some modern Windows 10 Computers. I see this as a serious problem in the very near future. Does anyone else share my opinion?
NI is very likely looking at options for replacing Silverlight to manage the roboRIO. That said, IE is available on every supported version of Windows, including all versions of Windows 10.
Can you explain the security issues that you’re specifically concerned about relating to using IE and the Silverlight plugin to access the roboRIO? I’d be happy to connect you to the right people to report them. I’m quite familiar with the security bug process inside Microsoft and can assure you that they’re taken seriously.
That’s a similar thread I made, it has some useful workaround tips in the comments.
Do. Not. Hold your breath. Waiting for this. Most mentors can recall situations where a vendor was late with a patch, or the patch didn’t work, or they couldn’t deploy the patch because some other dependency needed updating.
We’re talking about a ground-up reimplementation of NI’s config tool. Even if NI completes it before next January, FIRST might not be able to deploy. Microsoft declared Silverlight end-of-life in 2013. I’d argue NI should’ve had this done in 2014, yet here we are in '17 with no replacement in hand. This isn’t a favorable sign.
Keep your browsers updated, and know how to reenable the SL plugin.
What is your source for this statement? Silverlight 5 shows as being on mainstream support until 2021, according to this official Microsoft statement.
It always comes down to a knowledge thing. What matters is that NI has something they feel they can maintain with this web interface. These guys wrote their own browser plugin to flash their offboard processor, but the people in charge of the firmware would probably rather keep up with their current code base than re-write it from scratch.
A junior developer may think “why not just stick then entire thing on a Rio-local node stack?”. Yet given how web interfaces usually try to resolve their own external links*, that makes a fundamental assumption. We can’t presume that the RIO will get internet access out-of-the-box given the kludgey circumstances of school wifi. So I’m not so sure replacing the web interface an easy feat. If a Flash, Silverlight, or a browser plugin deals with that then it’s likely the single-most valuable reason to keep the core code base around.
I hope to make you guys aware of the underpinnings of why things are as they are. Most of the time well-intentioned engineers find one of the only solutions to a core critical problem, and then design an entire system around that solution since finding a different solution is so time-consuming. In our case, it’s the fact that the robot doesn’t go to the internet by default.
Sure, local repos can be setup. Yet that takes time, and (usually) software installs on the local computer. The point is, this isn’t a simple “replace X with Y” problem - it’s an integration-of-unknown-environments problem.
If their own decision not to support it in Edge isn’t enough of a source, every announcement that MS is abandoning its use in their own products should be. These just aren’t the actions you take with a product you are actively developing.
Yes, it’s in “mainstream support”… meaning we will get bugfixes and security patches without having to pay for them until 2021. But it isn’t like new systems in 2020 will ship with a version of Windows that Internet Explorer 11 is supported on.