A Comment from "Where Is Smalltalk Weak?"

Previous Topic Next Topic
classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

A Comment from "Where Is Smalltalk Weak?"

Someone posted this at our Google+ page:

Firstly, Smalltalk MT had its last release in 2009.  Redline Smalltalk had its last commit (merged) in October 2013.  For practical purposes, they are dead!

Secondly, I shall table two big weaknesses from my personal experience.

1. Lack of a universally-implemented standard

Each Smalltalk implementation is its own silo.  The risk of a vendor providing poor support or (worse) going bust can be devastating.

Over a decade ago, I was on the verge of choosing Smalltalk MT for implementing a new product.  I did not because the initial prototype revealed several integration issues with third-party Windows 32 libraries.

Some Smalltalker from Europe then suggested Dolphin.  Over the following weeks, my team discovered how painful it was to even try `porting' the prototype over to Dolphin Smalltalk.

That killed all prospects of Smalltalk even reaching the upper management for an approval.  My own team no longer wanted to have anything to do with it.  The risk of getting stuck in a silo was very high!

2. Performance

I developed several applications for desktops.  In simple CRUD applications, sure, there is no performance problem.

But, even in mildly complex reports and visualisations, we encountered terrible lags when displaying interactive graphs, etc.  In addition, these lags would manifest more terribly because user input events were never captured reliably by the controllers.

The users were left guessing if their actions were understood by the system, and therefore they had to wait for a few seconds, or if they had to re-try the actions because the system lost the input events.

As the article acknowledges, we - of course - don't even talk about number crunching.


The above two were my biggest problems back when I used Smalltalk.

So I had to respond:

I've been in constant contact with Redline author, James Ladd. He is working furiously to get Version 1.0 out the door by the end of summer, so the project is definitely NOT dead!

Smalltalk MT's latest 64-bit release (Version 6.3) is dated February 2, 2015, so this project is also NOT dead.

No doubt, there is a theoretical risk to having siloed implementations, but you've overstated the risk. The "main" Smalltalk vendors (Cincom, Instantiations, GemTalk) are doing well and Cincom itself has been around for decades.

Other dynamic languages such as Python, Ruby, and Clojure are doing well as open source. Smalltalk has Pharo, which is the centrepiece of our campaign. So choosing open source could mitigate or eliminate the risk.

It's true that in the past, hardware has constrained Smalltalk's performance, but this was also true of other dynamic languages. Today's hardware is so good that no one complains of performance, except for special situations. In general, Smalltalk's performance is on par with Python, Ruby, Perl, PHP, and Lua, all of which have their ardent followers. (See http://benchmarksgame.alioth.debian.org/.)

So, yes, choose the right tool for the job. Smalltalk is not good at everything, but it's more than good enough for most conventional applications.

You received this message because you are subscribed to the Google Groups "Smalltalk Research" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.