Ask HN: What should Microsoft do to change geek opinion?
Since (and even before) Satya Nadella become CEO of Microsoft, a lot has changed. From adoption of Open Source technologies to open sourcing a big chunk of its active software stack, from hosting Linux on Azure to supporting Office and .NET on all major platforms, not to mention its active contributing to Hadoop and a lot of other projects.
Having said that all, I can see little being changed from the attitude of the community towards Microsoft. Some are still skeptic about how real a move it is while others feel it is too little too late. Some (including me) feel this comes out of utter desperation rather than a strategic move from a change in the vision of the company. And ss Satya himself said, the biggest challenge is changing the internal culture.
So would you think attitude of the general geek community has changed towards Microsoft? And if not, why that is and what would make you change your opinion?
22 comments
[ 3.2 ms ] story [ 71.7 ms ] threadWhat file system layout would you want? UNIX? Does anybody think the current way the UNIX FS is used makes any rational sense? It is a giant disjointed mess, with a great deal of misusage by developers and users alike.
Microsoft could dispose of e.g. Program Files, Program Files (x86), Common Files, and also re-organise %AppData% and %windir% a decent amount. Remounting on / wouldn't provide much benefit (and technically Windows NT actually is mounted at \ they just use the root context for object storage, not files).
They have been trying some re-organisation with the Windows 8 "app" system. But since that has not been very successful neither has the project seen many results. Essentially they're siloing apps in the same way mobile devices do e.g. iOS apps/Android apps, etc.
SVC already is completely driven from the command line already (there is no UI).
> In general for TCP/IP there is still just almost nothing.
No clue what that means or what it is in reference to.
I don't want powershell, I want the things I'm used to. For example, Mac OS gives me that.
Unfortunately I believe the only thing that will change opinions are a considerable amount of time and effort in the other direction.
cron is missing a bunch of Task Scheduler's toys e.g.
- Start task on computer idle (and no, "nice"/low priority is not what this does)
- Stop task if the computer ceases to be idle, restart task if idle state resumes
- Start only on AC, stop if computer moves to battery
- Wake computer to run this task (unreliable, but useful when it works)
- Start task only if network connection is available (<pick network connection>)
The following has to be hacked with a bash script:
- Random delayed start
- Stop tasks if they run too long
- Begin on event (list of 300+ events, +custom event filters)
Honestly your criticism is pretty silly. cron is very very basic, and has to be hacked to do most things. Task Scheduler is incredibly powerful and doesn't require giant ugly hacks or custom scripts to accomplish things. I really think your reply is born out of ignorance (of both cron AND task scheduler) or just a raw unwillingness to learn something different.
Microsoft COULD put cron into Windows, just remove 70%+ of task scheduler's functionality, and all knowledge of how the underlying system works. But at least it would be cross platform...
I don't have numbers to say "a majority", but a whole lot of developers don't use Windows. They really need to give up on the Windows-everywhere stuff, but as I understand it that was a Ballmer thing. Anyway, I mean, they're Microsoft. I'm sure they'll land on their feet.
Hire cool agencies like Hello Monday (they've done most of Google and Android stuff).
Mimic the Unix command line environment of Mac that runs the various untilies (eg, homebrew).
* Ship native SSH (inc scp/sftp) and rsync clients and servers with Windows to make interoperability easier in heterogeneous environments. eg make SSH a first class transport option for Powershell remoting.
* Simplify the server licensing complexity, and then stop changing it all the time.
* Stop the secure boot shenanigans.
* Stop treating most HA, security or robustness features in SQL Server as "value adds" that are only in the really expensive editions.
* Contribute more engineering resources to improve open source interoperability projects and various 'DevOps' tools eg WinRM libraries, Configuration Management tools, Packer, Vagrant etc.