Alan Grimes escreveu: 
For me the biggest issue has been trying to run my code from outside
Squeak. This includes running Squeak headless to do something
script-like and configuring a GUI application to run in a way that
doesn't require the user to know they are running Squeak. Both of these
are supposedly possible, but very difficult to get right.
    

1. The VM, it's weak, no multithreading, few IO options.

2. The restriction to image based smalltalk instead of the ability to
run discreet programs...

Image based smalltalk is awesome but it makes it difficult to interface
smalltalk code with external systems.
The shortcomings about the squeak VM can be easily surpassed. Same thing about the "image" issue.

IMHO what holds Smalltalk (and squeak in particular) is lack of investment ($$$) in order to provide the things that "commercial users" (aka "regular developers/users") imagine/request as desirable (like "canonic" documentation, "canonic" books like "Smalltalk for dummies" (LOL) or "coreSmalltalk" or "Smalltalk Foundation Classes", better default options for the interfaces).

Besides these "small things" there are some important issues: there are several flavors of smalltalk and they're not compatible to each other. It seems that in the "open/free world" squeak will take the lead and perhaps it is good news. Anyways, today we have only VW to be seriously considered in the "commercial world" (since Dolphin is not multi-platform and was "half abandoned" by its developers and other smalltalks are both non portable and little known) and squeak and VW VMs are not compatible (no instant or even easy port from squeak to VW).

It is not possible to imagine the success of anything that is not accepted by the non-academic community. Currently most of the non-academic community in the World has little more than high-school degree, little or no fluency in English and earn something like US$1.000,00/month (or less) for a journey of at  least 40hours/week (and lots of non payed over-time). So I do think that the two previous paragraphs are relevant aspects to the question of "what is holding back smalltalk?"