Re: The problem as I see it
23 posts • joined 29 Jan 2016
EF was always handy for prototyping, EF Core is - though I personally am still suspicious about it in many cases - pretty much production ready. As long as you you understand what it's doing, how expression trees work and you're comfortable with ExecuteSqlRaw(sqlstring) when necessary. It's come a long way, although I admit that the singleton thread-unsafe context is infuriating at times.
However, I strong agree with the article that making every last ASP.NET sample, document and tutorial use EF Core is unnecessary, a barrier to entry where you don't know it or your shop doesn't use it and smacks of upselling, which is distasteful.
B- needs work.
There is profitable, and then there is profitable. How long do you think it'll take GitHub to return the 7.5 billion dollars that Redmond spent on it? My guess is never. When you think about it, MS didn't actually purchase anything of value THAT THEY CAN RETAIN.
In absolute terms, MS didn't spend 7.5 billion dollars on it, not really. It's an entirely stock-based transaction at today's valuation (which is high). Now sure, they could have sold that stock instead but that would depress the market. Even so, I agree that they will almost certainly never gain 7.5bn worth of profit from GitHub but what they can gain - not necessarily will but can, depending on how they manage it - is twofold.
1. They become officially the Open Source Company. You can expect more of MS's stuff to go straight to GitHub from now on, including their IoT linux implementation and almost certainly Windows Core. Everyone here has been telling us all that Windows is doomed since Vista, hasn't happened and probably won't but it could regain some of its power that way, ironically enough. Their other software would also benefit from this approach in many ways and Nadella knows it. If he isn't sure, Scott Guthrie certainly is.
2. The competition in the tech industry right now is not AIor machine learning or even cloud computing. It's all about who has the best people. Add GitHub to LinkedIn and MS have an eagle-eye view of who's on top of their game in any given field and that, well, that probably is worth billions.
It's entirely possible, jake. I know you hate Microsoft on principle but they do have Enterprise sales channels that GitHub certainly doesn't. In fact, GitHub doesn't need sales staff at all now. Or HR or accounting or payroll. And of course, GitHub is on AWS right now but I expect it'll soon be moved to Azure, where it can bask in free hosting and free bandwidth.
So all of GitHub's costs just went away.
Yeah, I expect it could be profitable. What do you think?
and you've no problem anymore, yay penguins yay.
Except - if I were to try to move my stuff off GitHub it woulld probably take three to four weeks, what with the outstanding pull requests. Those take time to go through in genuinely open source projects.
If you've no problem anymore - not that there was a problem outside your own head anyway - then either a) you're coding alone or b) you weren't doing anything worth doing.
Honestly, what did you bunch of howling support-droids want to happen?
GitHub raised $100m in the first round of funding, $200m in the second round. It was losing $22m per quarter with no hope of an uptick. When you owe VCs $300,000,000 and you lose $88,000,000 per year, the only hope is acquisition. Otherwise you go broke.
So somebody had to buy it. Who did you want to buy it? Amazon? Well, it's an option but to judge from Amazon Lumberyard and other products, you could forget about deploying anywhere except AWS if they did. Did you want Google to buy it? Do No Evil? Remember, this is a company that actively closes source on its biggest projects (Android), shitcans software at an unbelievable rate and, oh yeah, reads your damn email.
Who else? Facebook? You think they're the good guys? How about Oracle?
Oracle, jesus jumping christ on a unicycle. And you're whining about Microsoft?
Where's .NET? Oh, it's on GitHub. Where's SignalR? Oh, Github. Where's Blazor? Oh look, Github. Where's every recent MS codebase? Where's VS Code? Who writes SQL Server for linux and Visual Studio for Mac? Along with a billion other open source projects and, oh yes, a linux distro.
"muh muh muh evil muh muh convicted monopolist muh Ballmer muh is it still 1992?"
No, it isn't. This was the best choice for GitHub. Suck it up.
just to cover Farmer John who won't pay for a broadband line and a pico-cell.
Where is Farmer John supposed to buy a pico-cell? Nobody sells any that you can legally connect to a UK network any more. Vodafone claim to but refuse unless you're a business account with a minimum of 20 numbers. Three don't. BT/EE and O2 just say "use our app" which is a) crap and b) only works inside your wifi zone, if at all, and very badly even then.
Everyone understands that rural coverage is hard - although apparently not a problem in rural India or Africa where mobile payments are enormous - but let's not blame the people who live there for that.
I don't much care. From the dev/architect point of view, what I want from a BA is basically a single UML data flow diagram and a single UML State Machine. And even the State Machine is fairly optional.
Give me those and I'll design and build you your system. Give me a bunch of HTML "wireframes" that have been "tested" and I walk.
Imagine for a moment that you're tasked with managing a replacement process for one that's failing badly. The business can't keep anyone doing this - essential - job for more than two months because the process is so fucking awkward that they quit.
Let's assume the current process is built of a big set of data, held in Oracle and a few other places, manipulated by spreadsheets, published to the whole company and then processed. Let's say it affects employee pay, so it's not something you can afford to get wrong.
The correct way to do this would be to use a BA. Hire a decent business process analyst and sort out how it should work. What you can't do - but usually happens - is that somebody goes and asks the users what they want.
What the users want is inevitably going to be what they've already got plus magic that makes it less horrible. This is not a well-defined requirement and we already know the process is fucked up (and almost certainly hard for even a BA to analyse since the users won't know it well because they are by definition new to the job, what with nobody lasting longer than 8 weeks).
The worst case scenario is letting UX talk to the users before the process is set in stone because then you'll get what they've already got with jQuery and extra magic whereby impossible data is displayed instantly and billions of calulations are performed to fill in every read-ahead dropdown.
Developers will then inform UX that what they've ordered is impossible, UX will hurl all toys forcefully from the pram and insist "we've tested this", devs will respond by demanding to see the test conditions and metrics, UX will patronisingly explain that "testing" means "showing it to the users and they like it so now you have to make it" whereupon any decent set of devs will proceed by utterly ignoring UX and making their best guess at what works. Which will work but nobody will like it.
I'm not evangelising BAs here. There are good BAs, bad BAs and utterly incompetent BAs. But you really need the process sorted before you even take a step or your end result is guaranteed to be bad. And yes, devs lack soft skills that UX - for example - don't. But there is no substitute for defining the process exactly, even if design and even feature-set remain agile and fluid.
Stewardess - we have a medical emergency! Is there a doctor on the flight?
Vegan - I'm a vegan.
Q: How can you tell when someone has an iPhone?
A: They tell you.
And finally, bash in Win10 means you finally got the year of linux on the desktop. And millions of users are running it. And they all paid Microsoft for it.
How do you feel. Tell me now, how do you feel.
I'd love to know why the Samsung SCX-3200 at home refuses to print anything at all when plugged into the desktop PC it sits next to, but if you unplug it (involving moving furniture and delving through electrical spaghetti) and plug it into a laptop, it works first time.
I've tried swapping out cables, changing USB ports (which work perfectly well with every other device), kicking it...
Biting the hand that feeds IT © 1998–2020