Archive for September, 2017

as expected

Saturday, 30th September, 2017

and the actual message:

1,000 posts, 10 years (and VBA still going strong).


Is VBA a dead end ?

Thursday, 28th September, 2017

This topic came up on Excel-l, I got shouted down on there, so I thought I would elaborate on my views here. (its mainly an Excel view but I don’t think the other MS Office apps are much different.)

I see two sides to the VBA story – the technology and the career.

Technology wise a VBA solution is a dead end – it can’t be run on a server (officially), it can’t be run in a browser, it can’t be run in a high performance cluster, it can’t be run on mobile devices. This means it can’t be easily scaled if more users or geographies must be served. You can’t really even have multiple simultaneous users of an Excel VBA app. With the relentless march to cloud and mobile, VBA is left behind.

Of course MS won’t remove support for VBA, the world would fall apart, well the finance, pharma, and energy industries at a minimum. But the editor is pitiful by the standards of this century. It is testament to the work of those MS guys in the ’90’s that the debug Edit/continue is still better than nearly all mainstream IDEs. But its still out of date and hard to extend.

MS can’t invest in VBA because their strategy is .net/cloud/browser blah blah. They had the chance to fight on their strengths – rich client, but instead they tried to be google. And lost. Now we are all struggling with crap web sites in bloated browsers with slow Win3.11 style interactions from 1995.

VBA hasn’t suddenly become bad, its still very good at what it does, its just that we are often now looking for more from our solutions. More users, more device types, more threads, more whatever.

So its not dead but it is basically limited to single user apps on their individual workstations. But I am sure it will continue to work for a long time.

From a career point of view, what else is like VBA? nothing really, so expertise in VBA doesn’t easily lead onto anything else, hence its a bit of a dead end.

VBA is not really object oriented so it doesn’t lead to the sort of object based designs that C# and Java are suitable for. Its not functional so it doesn’t lead to F#/Scala.

Of course you can write VBA in classes, but lack of implementation inheritance hobbles your solution. By the same token you can write VB style ad-hoc procedural code in C#/, but you would be missing out on many of the elegant OO features of .net. You would struggle to implement most modern design patterns in VBA, and the latest CV must-have MVC is also all but impossible.

From a functional point of view (probably more useful than OO these days) VBA can’t treat functions as data smoothly (although CallByName does better than many languages). And a little toot of Application.Run can ease the friction too.

In Excel VBA especially we can get an awful long way without worrying about user interface/user experience or data structure. This is great for your current users (fast development/deployment) not so good for your career development – no other system is so well integrated as Excel/VBA. No worrying about Restful APIs, onClickListeners, Asynchronous callbacks, all very powerful, all generally avoided in VBA land.

If you love VBA, and why not, then continue on, you are probably adding untold business value with the work you are doing (and perhaps a little future maintenance pressure for someone). But I’m not sure it will take your development career too far, business career possibly but IMO it won’t help you much as a developer.

If you are doing VBA now and thinking of progressing to other development technologies then consider carefully how you will make your next steps. Adding Access and SQL to VBA is completely doable, adding modern scripting, web technologies or .net/Java is much tougher.

Getting into .net via is a possibility, but, I think it would be easier to go straight to C#, although the .net/Excel story is still, in 2017 very ugly.

ExcelDNA might be a good bet if you want to inject a bit of .net into your Excel life. I found this approach much more viable than VSTO infrastructure hell. Somebody somewhere is maintaining a lot of both that I developed. I bet the ExcelDNA maintainer is having more fun.

I also managed to blag my way into writing a load of custom Java on Essbase, that route probably isn’t open for many though.

Which led to my new focus on Android which is both challenging and rewarding. Although stay posted for news on that…

btw this is my 1,000th post on SOS, I assume WordPress will give me some superficial badge of honour, reward or ‘achievement unlocked’ bollocks, in the modern way.




Job hunt

Wednesday, 20th September, 2017

I’m back on the hunt for a contract.

Looking for easy work, high pay, and ability to work from home in my pyjamas.

I am just caught mid hop really, my Android mastery has not quite reached the level to dazzle, and my Excel/VBA/trading experience is starting to look a little distant.

That said I do have a few mini projects on the go both in VBA and Android, but a big juicy contract would be handy enough too

I am pimping myself out on all the usual job sites and I have noticed a pattern.

An agency posts a job, I apply with my cv, and then… tumbleweed.

This is not new of course, and perhaps my ‘varied’ cv is not what they are looking for but…

A couple of agents did get back to me either with a no thanks or a discussion or whatever. And you know what?

The next time a role comes on with multiple agencies, I apply to the ones that bothered their arse to get back to me. Those tumbleweed guys? I don’t bother.

So pimps of the world, if you can’t even be arsed to ping back a quick no thanks to your applicants, you may be losing out on candidates in the future. just sayin.