The efforts of Uber and Lyft over the last few months to rewrite the laws of the City of Austin to their pleasing through a not-at-all-subtle application of money, spam and propaganda went far, far beyond taste. Doesn’t matter if you support the TNC companies or not (I mostly do), but they went way too far in their blatant attempts to subvert the democratic process of our city. So, yeah, no thanks – don’t let the door hit your ass on the way out.
I imagine they will have better luck corrupting the State of Texas legislature to override the cities, but have fun doing this 49 more times.
According to Mike Harsh, a beta version is coming in an insider release "after the Build conference." [1] During an interview Q&A, Kevin Gallo says that it might show up in a week or two. [2]
No, that's a fair interpretation of my comment - after all, the meaning of "cloud optimized" is kind of vague. However, if you compare the new version and previous version and the new version ticks more positive boxes for cloud usage scenarios, then it's not unreasonable to call the new version "cloud optimized".
Beyond the modularization efforts, one could look at the Linux port and dramatic performance improvements in the same light.
Actually, I don't see it as pure marketing fluff to call the new ASP and .NET Core "cloud optimized". Breaking the framework into smaller pay-for-play components to reduce deployment footprint and allowing for xcopy deployment of self-contained binaries without a globally installed .NET runtime really does help enable new scenarios for cloud containers and microservices.
It's a shame, since these topics could make for a rich conversational interview that probably has more to do with the actual job than whiteboarding an algorithm. Just taking source control as an example, a hypothetical interview question could be something like:
"You've been assigned to implement feature X in our product. Assume that the specs are clear and you have a pretty good idea what code you are going to write. Also, assume that your teammates Joe and Jane have been assigned features Y and Z respectively with roughly the same due date as your feature. Ideally, describe how you would make your changes, test them and coordinate with your teammates to get all three features merged into the master source code."
Obviously a lot of details are missing, but they can be fleshed out in conversation (during which the interviewer should also describe the current process used on the team the candidate is being hired for). This would give both parties some insight into experience and expectations.
It seems like there is a large middle area of concepts in between low-level algorithms/data structures and high-level system architecture that are left out of many of these interview prep guides:
* Principles and patterns of object-oriented (or functional) design
* Relational (or NoSQL or analytics) database design
* Unit, integration and system testing
* Logging, profiling and debugging
* Source control (e.g. branch/PR/merge flows)
* Deployment and devops
Do these subjects really not come up in some programming interviews?
Not sure if they were working together with GitHub on this, but Microsoft also announced today that Visual Studio Online Git repos now support Git-LFS with unlimited free storage:
At least for Windows 8/10, I'm not sure if something like React Native is really necessary, since the platform itself already has built-in support for making "native" apps in HTML/JS.
There will be three more beta releases scheduled for July, August and September, followed by a release candidate in November. RC 1 will have a go-live license and support for production deployment on x86/x64 Windows, Mac and Linux.
There may be more than one RC release, depending on feedback, but the final release should be some time in Q1 2016.
Beyond retrieving the contents of several unrelated enum-like tables in one roundtrip, multiple result sets are also super-useful for grabbing a deep relationship tree of joined tables in one roundtrip.
Writing a single projected SELECT with a deep list of JOINs causes the single result set wire size to explode as you include more and more to-many relations. However, breaking up the joins into separate SELECT roundtrips to reduce wire size will increase latency.
With multiple result sets, you can take all the separate selects, stuff them up in a stored procedure, connect them with insert joins through temp tables, then return multiple result sets from the temp table contents. This gives the best of both worlds: low latency from one roundtrip and a small wire footprint.
I imagine they will have better luck corrupting the State of Texas legislature to override the cities, but have fun doing this 49 more times.