alternatives for AUTOLOAD

[fschwach]

Just wanted to raise this as a new issue because the pull request in which we discussed this originally is already closed but this issue is still open for discussion.

The problem: there is a large number of tags for input and output to/from primer3 for which we may want to create accessors. Writing all those accessors explicitly as methods will bloat the code quite a bit and may be a maintenance problem. AUTOLOAD is not an ideal solution and Bio::Root::Root::_set_from_args is only meant to create accessors from arguments given at object instantiation, which is not what we need.

I think these are the options:

1. We could auto-generate accessors for the various output tags of primers and primer pairs by setting up methods from a list of allowed tag names (similar to what I have already done in Tools::Run::Primer3Redux::_create_run_methods) but I'm not sure if that is acceptable style for BioPerl?
2. I can create all the subs for the different tag accessors and paste them into the code.
3. Instead of creating direct accessor methods for each tag by name (such as $primer->hairpin_th), we could have methods "has_tag" and 'tag_value' where the tag name would have to be given as an argument. Maybe that's the cleanest solution.

What do you think?

Frank

[JohnMCMa]

The easiest way is to apply option (3) by reading the different tags that already existed in the infile, and make an array of legal tags for that result/pair/primer etc. This way we can solve a lot of headache in writing sanity check codes.

I'd say, though, I made this comment on the output tags--the input tags is a can of worms that I would open another issue to discuss.