[BioC] Using htmlpage

[john seers (IFR)]

Hi Kasper

Thanks for the reply.

>This is exactly what namespaces were designed to avoid. You can  
>perhaps work around it with assignInNamespace, but it will be ugly and

>a hell to maintain.

I am not sure that is quite accurate. Namespaces are to avoid altering
the namespace code either accidentally or on purpose. If a hook is
supplied for a user function that does not break the idea. 

>You can perhaps work around it with assignInNamespace, but it will be
ugly and  
>a hell to maintain.

Exactly. I do not want to alter the namespace code or maintain it. Nor
do I want to work around it using the assignInNamespace. (Perhaps you
did not read the start of the thread?)


Regards


John
 
---

-----Original Message-----
From: Kasper Daniel Hansen [mailto:khansen at stat.berkeley.edu] 
Sent: 05 February 2009 16:56
To: john seers (IFR)
Cc: J.Oosting at lumc.nl; bioconductor at stat.math.ethz.ch
Subject: Re: [BioC] Using htmlpage


On Feb 5, 2009, at 3:41 , john seers (IFR) wrote:

>
> Hi Jan
>
> Thanks for the reply.
>
>> By using the triple colon you can use objects/functions that are not
>> exported from the namespace. i.e. annotate:::getQuery4Affy()
>
> That would help if I do not have to "assignInNamespace" each  
> function. I
> will try it. Thanks for the suggestion.
>
> Really my problem is the other way round though. I want the package to
> call my (user) function. Is there a currentworkspace:::userfunction()
> option?

This is exactly what namespaces were designed to avoid. You can  
perhaps work around it with assignInNamespace, but it will be ugly and  
a hell to maintain.

Kasper