thanks lukas On Aug 29, 2010, at 11:29 AM, Lukas Renggli wrote:
Some comments:
- The term "Widget" is not used in the architecture of the OB framework anywhere. Maybe it would be better to replace it with the term "Panel" that is used in the framework.
- Figure 1.1 depicts a bug of the un-themed scroll-bar that has been fixed a long time ago when Polymorph was introduced.
- p. 9 at the bottom there is a comma disconnected from the previous word.
- p. 12 "This enables a smooth gathering of independently realized commands." -> "..., and allows one to cleanly extend existing browsers without having to patch the system (e.g. refactoring browser)."
- p. 13 The definition has nothing to do with editable text. By overriding the right methods in the OBNode a node is directly editable as well. The reason to have a separate OBDefintion is only to be able have multiple different definitions for a single node (e.g. to provide different views on source code from an OBMethodNode: source code, byte code, decompiled code, pretty-printed code).
- p. 14 "Each of this method should return the class of the command (and not an instance of it)." This is not true. The requirement is that the returned object responds to #on:for:. There are examples where factory objects or OBCommand instances are returned.
- p. 15 "These actions enable interaction with the domain model, however they do not allow further navigation in the metagraph." This is not true. There are various ways to navigate and update the metagraph, e.g. aNode signalSelection (see p.20 "Widget notification").
- Figure 1.5 does depict O2, not the "OmniBrowser System Browser" that is described in the text in Section 1.4.
- p. 24: The acknowledgements fit to O2, but they seem distorted in the context of OmniBrowser.
Lukas
-- Lukas Renggli www.lukas-renggli.ch
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project