Re: [LAF-devel] Re: Design

[David M. B. <da...@da...>]

Thanks for starting to get goals into a list form.  I like good, concise 
lists. :)  I have a couple of comments here; let me know what you think.

Wesley J Landaker wrote:
> On Saturday 09 August 2003 11:03 pm, Samuel G. Shirley wrote:

[...]
>>>>I need Paf and GedCom compatability.  I think it would be great
>>>>if we can fork something to be PAF compatible, however, I'm not
>>>>overly optimistic about this.
>>
>>I agree with this. I've already tried a couple and I am not happy. I
>>tried Genes (genes.sourceforge.net) and Gramps
>>(gramps.sourceforge.net). Genes is not in a good working stage.
>>Gramps looks very pretty and works pretty well. But there are too
>>many quirks in it for me to be happy with it. Frankly, it's a pain. I
>>like PAF. I personally would like to see a clone of it in Linux. This
>>would also serve to make those moving from Win to Linux much happier.
> 
> 
> I haven't tried any lately (which is why I suggested that we look at the 
> state of things now), but back when I tried them, most geneology 
> programs were sorely lacking. 

Ok, so my original idea for LAF was to be a PAF clone.  This especially 
goes for both ends of the software usage (GUI and file format, the 
former being most important (as long as we don't get into trouble)).  I 
wanted a PAF user to sit down and start using LAF from day one and not 
even know they were using LAF.  Sure, there are going to be some subtle 
differences, but as much as is feasibly possible make it a PAF 
look-and-feel-alike.  That was the first and foremost feature and all 
other features, as long as they don't detract from this first feature, 
would be considered.

With that in mind...

> In fact, even between PAF versions there have been quite a few changes 
> in usability! I know people who still prefer PAF 3. =)
> 
> Anyway, it sounds like between everyone here we've got the following 
> goals: (Correct me if I'm wrong. ;)
>   - Be feature compatible with PAF. That is, support doing everything in 
> LAF that is doable in PAF.

Right on!

>   - Be file-format compatible with PAF. That is, support transferring 
> data with *no loss* between PAF and LAF. Whether this can be done with 
> GEDCOM or if it requires native file-format compatibility will have to 
> be determined.

Yup!

>   - Be familiar to PAF users. That is, our GUI should work as much like 
> PAF as makes sense.

Even a little into the "does that really make sense?" realm should be 
done.  In other words, there may be extra effort invovled (more than 
usual) to make it look like PAF so that PAF users are completely 
comfortable.

>   - Be *better* than PAF; be open and extendable. That is, if we can do 
> something better than PAF does it, or add new features that PAF lacks, 
> we should do it.

Yes, but only to the extent that we don't take away from the PAF 
similarities.  I'll admit there are things in PAF that I would have done 
differently, but there will be some things that we'll have to implement 
that we won't like just to make it familiar to PAF users.

>>>Question: how important is it to you to have PAF file-format=20
>>>compatibility? GEDCOM is certainly no problem, but the PAF
>>>database=20 files themselves have changed format in every version
>>>of PAF and I=20 don't know if they are documented anywhere...
>>>certainly we could=20 reverse-engineer them if it's worth it to.=20
>>
>>I think its important to have PAF compatibiilty. PAF is a big program
>>with a large following. Our program should be able to AT LEAST import
>>those files.
> 
> 
> I agree that importing PAF native files is important, *even if* GEDCOM 
> round-tripping is possible without data loss. However, this may (or may 
> not) require some reverse-engineering effort. We might want to start 
> off with GEDCOM support, but out file I/O subsystem should be 
> pluggable. =)

Exactly.

I'll start a list on SourceForge of the project goals as soon as I see a 
little more feedback on what we think the goals should be.  Your email 
is a wonderful start (and is probably close to the finished list) for 
everyone to comment on.  At least we have a few people now on the list 
who can comment (more than we had before), but the number is small 
enought that I'd hope everyone would comment.

Thanks,
David