On Sep 30, 2007, at 2:31 PM, Egon Willighagen wrote: > On 9/30/07, Rajarshi Guha <[EMAIL PROTECTED]> wrote: >> On Sep 30, 2007, at 2:26 AM, Egon Willighagen wrote: >>> On 9/29/07, Rajarshi Guha <[EMAIL PROTECTED]> wrote: >>> Why not simple WYSIWYG? -> You see no double bonds (SMILES, MDL), >>> but >>> marks of sp2 (SMILES) or aromaticity (MDL bond type 4), so you >>> get no >>> double bonds but markers of sp2 or aromaticity. >> >> That's fine - but you're going to have to put up a very big notice, >> preferably in flashing neon, reminding people that after SMILES >> parsing one must do bond deduction etc etc. > > OK, we can surely put in some <font style="color: neon"/> in the > JavaDoc and there used to be some <blink> in HTML too.
:) > Go on. What atom type list would you suggest? Maybe the CDK superset by default or else a user definable atom typing scheme. For certain formats, may be a specific scheme would be auto-selectd > I only know of Mol2 that > actually uses this approach. SMILES does not generally have this > approach, though valencies are defined for the organic subset. How do > you envision the 'setting'? As an argument to a file format class?) Ideally, I'd like to have some sort of global variable, so that any code used after setting the atom type scheme would be able to check what scheme was specified. It's not entirely clear how one would achieve this. > Why is that prior to the reading, instead > of during the reading? I suppose either would be fine - I just separated the steps out for discussion. As I said above, it could be a flag to the file reader classes > And why does this contradict the Lego > philosophy? Given Christoph's point about a file parser, just parsing files, the approach I described above would result in a lot of things going on all together, rather than each peice doing it's own job and leaving it to the caller to take the appropriate steps in the proper sequence to finally get a complete (bonds perceived, implicit H's evaluated etc) molecule BTW, on an unrelated note, is it possible to have the Javadocs include a link to the source for a given class in the SVN repo? It would be a handy way to navigate to sources, without having to build the URL to the SF SVN repo ------------------------------------------------------------------- Rajarshi Guha <[EMAIL PROTECTED]> GPG Fingerprint: 0CCA 8EE2 2EEB 25E2 AB04 06F7 1BB9 E634 9B87 56EE ------------------------------------------------------------------- "whois awk?", sed Grep. ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Cdk-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/cdk-user
