Off-topic software discussion

[The Author Guy]

I follow a similar motivations, I tend to focus on what my clients can manage effectively.

Most of them are small to mid-sized financial services firms, they have money to spend on IT, they just don’t have the inhouse knowledge they need (at the manager level) to do much, and they have no interest in a huge IT department.

Historically, that meant Windows as the solution most affordable to them, because it was so much easier (in the 90’s) to find reasonably priced windows system admins and developers, and you didn’t have to be super techy to screen these people, no arcane *nix info.

Linux was for the big guys, and of course, their DB options were limited to Sybase, Oracle and a few other ridiculously expensive options (assuming they wanted support) MS SQL Server 6.5 came along, it was cheap, had a GUI, anyone with a PC background, even a lay person could work with it and learn it pretty easily. (used to work for a person who thought every financial analyst needed to know SQL and have SQL Management Studio installed)

And it was Sybase in a end-user friendly form. And I gotta say, it was so much easier to manage quickly/out of the box compared to running Sybase on Suns, which I had been doing. Sybase at the time wanted dedicated partitions etc etc. and all sorts of contortions.

When I moved from academia, the business world was mostly Sun (in academia, SunOS/Solaris etc was only deemed suitable for file servers, for real servers you wanted SGI, IBM, HP, DEC (both unix and VMS) and I don’t remember the others ). However, buying Sun servers with Sybase, or Oracle and finding people that could manage it was very hard/expensive for a company with <50 employees.

Now, I've been doing PC's since the introduction of the IBM AT, before which I was mainframe/teletype or Apple II. Real computing started with Sun workstations, I laughed at the first Macs because they were so primitive compared to my Sun workstation or my X-terminal and the graphics were horrible.

I then moved into all real work on VAX, Amdahl, SGI, IBM and got serious about Big Data (such as it was circa 90-95). Linux 0.9x and BSD were fun toys that you could use as a workstation if you were adventurous (but because my team had really big budgets for building supercolliders, we got the good stuff, the expensive and fast stuff)

Then went to business, and no one had my old budget and linux was not yet ready for Prime Time. So Windows took off.

Nowdays, yes Linux is ready for my client base, for servers. Lots of nice/tested OS apps/platforms with well tested scripts, it's never been easier to have a Linux box. I'd argue this is at least as important as the price feature. Ease of use for the low to mid-tech people is critical. This wasn't there until within the last ten years.

But application development? The programming overhead is still a problem. It is still much easier/cheaper to get C# or VC++ people. Unless you are just doing web development.

Don't know if you've noticed, but every other job opening seems to want a Java Developer. Far more expensive to get good ones than C# developers. Same with C++; a truly good C++ developer is hard to find, but there are a lot of so so ones that end up costing you money in lost time, poor design. C# is extremely forgiving compared to every other major language, so even a mediocre developer can be productive in it for relatively low money.

Same reason Ruby/Rails and other prepackaged platforms are great for basic websites. Very easy to use..

Long diatribe....my goal is always to get something that my client can support.

Which is why I am one of the few people that goes in and talks people down off the MVC ledge. I've found a lot of peopled get talked into MVC by consultants, but they don't have the tech budget for long term support of MVC applications. There aren't enough MVC developers out there and it requires a much larger development team. Anyone can bang their way through webforms, Ruby, PHP, etc.

Now, all that being said. MS SQL is getting too big for its britches. PostgreSQL, MySQL, etc work reasonable on Windows, but better yet, download a custom tailored VPS to run any/all of those and be done with it. Nowdays there are great windows/macOS client tools, web tools etc for managing them.

I do think this is the future, it's just about how do we walk people there? MS has served it's purpose, [color=red](in particular to beat the crap out of even bigger vendors trying to lock you into even more expensive platforms: IBM, Sun, HP Sun, Oracle, SAP)[/color] but I think costs and end user quality will split the workstation market to Windows, macOS, maybe ChromeOS (???) and non-premium servers will be clusters of VPS or container servers, probably containers long run.

I think the high end will most continue to be high performance supported Open Source based products leveraging java for HADOOP and it's crazy bestiary. But that is still to be seen as to how exactly it will play out.

[Mikey]

[Mikey]

[quote]I am just not a huge java fan, I love the theory of it, I love so much of what it introduced, but I was just too burned early on by horrible performance and its inability to talk with non-java things. And what I considered a crappy visual interface. Now all of this has been fixed, but it was eaten by the forces of evil, a couple times over so…

On the other hand, I’ll still take it over C++ any day, I am in the camp that believes that the original C++ specification was a joke and that no one was ever supposed to actually use it. I mean, I did use it, I built a lot of stuff in Borland C++ well before VC++ existed but once I found Delphi 1.0 I said good bye to C++ and haven’t looked back unless forced to. Of course, I only use that for fun today, since very few people in the US use Delphi.

I was surprised to find Syncplify.me is written in Delphi. And it is a new, modern program. It’s a great SFTP/FTP/FTPS server. My first clue that it was delphi was it’s crappy help file. Delphi programmers never seem to have time to write documentation. [/quote]

Ever since people started actually listening to what I’m saying, my technology choices have been more about keeping the correct strategic doors open for my clients.

For example, the open nature of Java, starting first from the open-standards and implementation alliance fronts, and then graduating towards open source, has meant that for each product, there has been a viable competing alternative. This has meant that the real negotiating power of my customers, towards COTS vendors and contractors, started from a good place, and has grown ever since.

Not only that, but their investment in employee skills (also on the contractor side) has included a much greater ratio of general, reusable skills, such as language, POSIX, protocol standards to product-specific skills, than on the Microsoft side of things.

About twenty years ago Linux started to get more popular as a server operating system, culminating now to a over 90% market share, especially in the cloud space which is growing the fastest.

Now my clients are all heavily invested in using the cloud and developing cloud products. In that space, relying on Windows as your operating system would mean that you would be utterly in the mercy of a single vendor, which has a history of radical strategy changes at CEO departures.

Let’s see what the future holds. The current .NET Core development looks promising, I’ve even developed a Hello World .NET Core web application completely on Linux, using the provided Docker containers and the JetBrains Rider EAP. Rider should soon be in public beta, it’s already pretty reliable.

If you have time, take a look at Kotlin and TypeScript/ES2016.

[Author]

Posts