The current link to the source tarball is called "Package Source" hence the
quotes. Yes, I could check out the package using svn, but when browsing
through a Bioconductor workflow, there are these handy links to the package
pages that let me download and browse the source tarball without having to
type anything. I like the idea of replacing the source tarball link with a
link to the package source in svn.

Best,
Matt




On Tue, Jul 22, 2014 at 1:58 PM, Michael Lawrence <lawrence.mich...@gene.com
> wrote:

> Just check out from svn to get the source... way easier to keep up to
> date, and if you notice an issue, easier to make a patch.
>
>
>
>
> On Tue, Jul 22, 2014 at 10:49 AM, Matthew McCall <mcca...@gmail.com>
> wrote:
>
>> I hope the "package source" link is not on the proposed list of links to
>> remove. I often use these links to browse through the source code of
>> packages to learn from others' work. Also, it seems that making the source
>> code (even slightly) less accessible would go against the principle of open
>> source software.
>>
>> Best,
>> Matt
>>
>>
>>
>> On Tue, Jul 22, 2014 at 1:22 PM, Michael Lawrence <
>> lawrence.mich...@gene.com> wrote:
>>
>>> Agreed. Disabling the links is a good idea. There's really no good reason
>>> for someone to install packages manually. If a user really wants to mix
>>> release/devel, it is still technically possible but this change would
>>> strongly discourage it.
>>>
>>> For ensuring the user notices that a page is for the devel version , I'm
>>> still in favor of the simple notification box. Probably without the
>>> option
>>> to hide forever.
>>>
>>>
>>> On Tue, Jul 22, 2014 at 9:26 AM, James W. MacDonald <jmac...@uw.edu>
>>> wrote:
>>>
>>> > Given that we have an ongoing problem with people inadvisedly clicking
>>> and
>>> > installing things, is there a good rationale for allowing them to do
>>> so?
>>> >
>>> > Perhaps the landing page for each package should be stripped of links
>>> and
>>> > replaced with some indication of the availability for each package on
>>> the
>>> > various operating systems. There could also be a note indicating that
>>> > people can install using biocLite().
>>> >
>>> > Best,
>>> >
>>> > Jim
>>> >
>>> >
>>> >
>>> >
>>> > On 7/22/2014 11:48 AM, Dan Tenenbaum wrote:
>>> >
>>> >> Seems like there are two problems, first that the release and devel
>>> pages
>>> >> look similar, but more importantly that people are downloading and
>>> >> installing from the package pages when they should be using
>>> biocLite().
>>> >>
>>> >> I am open to the suggestions for making the release/devel pages look
>>> more
>>> >> different from each other, but I think something needs to be done
>>> about the
>>> >> second problem as well. Perhaps a popup that comes up when you click
>>> on a
>>> >> package tarball saying "The best way to install this is with
>>> biocLite();
>>> >> are you sure you want to download it?"
>>> >>
>>> >> Whatever we do probably won't happen until after BioC2014.
>>> >>
>>> >> Dan
>>> >>
>>> >>
>>> >> ----- Original Message -----
>>> >>
>>> >>> From: "Julian Gehring" <julian.gehr...@embl.de>
>>> >>> To: "Hervé Pagès" <hpa...@fhcrc.org>, "Michael Lawrence" <
>>> >>> lawrence.mich...@gene.com>, "Vincent Carey"
>>> >>> <st...@channing.harvard.edu>
>>> >>> Cc: bioc-devel@r-project.org
>>> >>> Sent: Tuesday, July 22, 2014 8:07:29 AM
>>> >>> Subject: Re: [Bioc-devel] Distinction between release and devel
>>> package
>>> >>> websites
>>> >>>
>>> >>> Hi,
>>> >>>
>>> >>> Tooltips that appear while hovering over selected links are easy to
>>> >>> miss.  This alone will likely not be clear enough.  We should convey
>>> >>> the
>>> >>> information that the entire website presents a different version of
>>> >>> the
>>> >>> package.
>>> >>>
>>> >>> The idea of a notification box that can be made visible by the
>>> >>> individual user seems tempting.  One can combine this with an
>>> >>> optional
>>> >>> cookie, to remember the state between browser sessions.
>>> >>>
>>> >>> Changing the layout of the devel page itself will also be helpful to
>>> >>> make the distinction more pronounced.  Hopefully we could approach
>>> >>> this
>>> >>> in a way that maintains the nice design of the bioc website.
>>> >>>
>>> >>> Best
>>> >>> Julian
>>> >>>
>>> >>>
>>> >>> On 21.07.2014 21:50, Hervé Pagès wrote:
>>> >>>
>>> >>>> Hi,
>>> >>>>
>>> >>>> In addition to these suggestions, how about using a special
>>> >>>> background
>>> >>>> color for package landing pages in devel?
>>> >>>>
>>> >>>> Cheers,
>>> >>>> H.
>>> >>>>
>>> >>>> On 07/21/2014 07:32 PM, Michael Lawrence wrote:
>>> >>>>
>>> >>>>> Or an unobtrusive "notification box" that drops down from the top
>>> >>>>> of the
>>> >>>>> page, saying something like "this is devel"; there would be a
>>> >>>>> dismiss
>>> >>>>> button and a checkbox for whether to show again. The user is free
>>> >>>>> to
>>> >>>>> simply
>>> >>>>> ignore it and proceed as normal.
>>> >>>>>
>>> >>>>>
>>> >>>>>
>>> >>>>>
>>> >>>>> On Mon, Jul 21, 2014 at 7:10 PM, Vincent Carey
>>> >>>>> <st...@channing.harvard.edu>
>>> >>>>> wrote:
>>> >>>>>
>>> >>>>>  how about a tooltip that reads "installation via biocLite() is
>>> >>>>>> the
>>> >>>>>> recommended approach to Bioconductor software
>>> >>>>>> acquisition, other approaches may lead to inconsistent
>>> >>>>>> package-sets"
>>> >>>>>> that
>>> >>>>>> appears when a reader hovers over a tarball.  i would imagine
>>> >>>>>> that
>>> >>>>>> this is
>>> >>>>>> how the "wrong package" gets installed, by manually using an
>>> >>>>>> inappropriate
>>> >>>>>> tarball.
>>> >>>>>>
>>> >>>>>> wrong documentation is not so easy but the doc on the devel
>>> >>>>>> branch might
>>> >>>>>> have a different tooltip cautioning the readers to be sure they
>>> >>>>>> want to
>>> >>>>>> read the doc on the devel version.
>>> >>>>>>
>>> >>>>>>
>>> >>>>>>
>>> >>>>>> On Mon, Jul 21, 2014 at 9:39 PM, Julian Gehring
>>> >>>>>> <julian.gehr...@embl.de>
>>> >>>>>> wrote:
>>> >>>>>>
>>> >>>>>>  Hi,
>>> >>>>>>>
>>> >>>>>>> Can we make the package websites for the devel and release
>>> >>>>>>> version of a
>>> >>>>>>> package more distinguishable?
>>> >>>>>>>
>>> >>>>>>> To elaborate on this: In the past, I have seen several users
>>> >>>>>>> having
>>> >>>>>>> problems with using bioconductor because they ended up on the
>>> >>>>>>> wrong
>>> >>>>>>> page
>>> >>>>>>> (mostly the devel page when they would have needed the release).
>>> >>>>>>>   This
>>> >>>>>>> resulted in getting the wrong documentation or installing the
>>> >>>>>>> wrong
>>> >>>>>>> package.  The pages are well designed, and there is no reason to
>>> >>>>>>> change
>>> >>>>>>> this.  However, the websites for the devel and release version
>>> >>>>>>> of a
>>> >>>>>>>
>>> >>>>>> package
>>> >>>>>>
>>> >>>>>>> look almost identical, and that these two get confused seems to
>>> >>>>>>> happen to
>>> >>>>>>> many users (me included).
>>> >>>>>>>
>>> >>>>>>> If you search for a package within the bioc website, the release
>>> >>>>>>> version
>>> >>>>>>> always comes first in the search results.  If you are coming
>>> >>>>>>> from the
>>> >>>>>>> outside (e.g. google), this may not be the case.  In fact,
>>> >>>>>>> googling
>>> >>>>>>> a few
>>> >>>>>>> packages names often returned only the devel page in the top 10
>>> >>>>>>> search
>>> >>>>>>> results.
>>> >>>>>>>
>>> >>>>>>> What are the feelings regarding this? We could add a header
>>> >>>>>>> section on
>>> >>>>>>>
>>> >>>>>> the
>>> >>>>>>
>>> >>>>>>> devel page that states that this is an unstable version not
>>> >>>>>>> meant to be
>>> >>>>>>> used in production settings, and provide a link to the
>>> >>>>>>> respective
>>> >>>>>>> release
>>> >>>>>>> version?
>>> >>>>>>>
>>> >>>>>>> Best wishes
>>> >>>>>>> Julian
>>> >>>>>>>
>>> >>>>>>> _______________________________________________
>>> >>>>>>> Bioc-devel@r-project.org mailing list
>>> >>>>>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>> >>>>>>>
>>> >>>>>>>
>>> >>>>>>           [[alternative HTML version deleted]]
>>> >>>>>>
>>> >>>>>> _______________________________________________
>>> >>>>>> Bioc-devel@r-project.org mailing list
>>> >>>>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>> >>>>>>
>>> >>>>>>
>>> >>>>>      [[alternative HTML version deleted]]
>>> >>>>>
>>> >>>>> _______________________________________________
>>> >>>>> Bioc-devel@r-project.org mailing list
>>> >>>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>> >>>>>
>>> >>>>>
>>> >>>>
>>> >>> _______________________________________________
>>> >>> Bioc-devel@r-project.org mailing list
>>> >>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>> >>>
>>> >>>
>>> >> _______________________________________________
>>> >> Bioc-devel@r-project.org mailing list
>>> >> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>> >>
>>> >>
>>> > --
>>> > James W. MacDonald, M.S.
>>> > Biostatistician
>>> > University of Washington
>>> > Environmental and Occupational Health Sciences
>>> > 4225 Roosevelt Way NE, # 100
>>> > Seattle WA 98105-6099
>>> >
>>>
>>>         [[alternative HTML version deleted]]
>>>
>>>
>>> _______________________________________________
>>> Bioc-devel@r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>>
>>>
>>
>>
>> --
>> Matthew N McCall, PhD
>> 112 Arvine Heights
>> Rochester, NY 14611
>> Cell: 202-222-5880
>>
>>
>


-- 
Matthew N McCall, PhD
112 Arvine Heights
Rochester, NY 14611
Cell: 202-222-5880

        [[alternative HTML version deleted]]

_______________________________________________
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel

Reply via email to