Alright, I think you are right. The possible confusion is not worth the benefits. I have switched the ordering back to the classic sorting. I will now look into the dataset selector and see how to revise it. Thanks for pointing these things out.

Thanks,
Sam

On Thu, Nov 6, 2014 at 8:18 AM, Peter Cock <p.j.a.cock@googlemail.com> wrote:
Hi Sam,

I found the old approach (new repeat blocks at bottom) worked fine
when adding blocks one by one and completing them as I went
(which means once you have filled in the new block, you have
scrolled down to the button ready to add another block if needed).

If find the new approach (new blocks inserted at top) visually
confusing as I am used to filling in forms from top to bottom.

I do concede the approach makes it easy to add several repeats
with a few clicks, and then fill them in - but personally that isn't
a common thing for me to do, and I do not think this change is
worth the confusion.

There are a number of other visual changes on galaxy-central
compared to the current release, so of which seem harmless
like the boolean parameters becoming a yes/no toggle in place
of a tick box, others are more daunting/scary (e.g. collection
related changes to the file picker for automatic batch jobs).

Is there any draft documentation on these changes being prepared
to go into the next release notes? This is the sort of thing local
Galaxy Admins would appreciate to anticipate local user queries.

Thanks,

Peter

On Thu, Nov 6, 2014 at 12:45 PM, Aysam Guerler <aysam.guerler@gmail.com> wrote:
> Hi Peter,
>
> Yes we changed this on purpose, however it is open to discussion. The
> advantage is that the user does not have to scroll down after adding a new
> repeat block. Additionally it enables users to easily add more than one
> repeat block quickly, since the insert button does not relocate on the
> screen after adding new repeat blocks.
>
> Thanks,
> Sam
>
> On Mon, Nov 3, 2014 at 7:25 AM, Peter Cock <p.j.a.cock@googlemail.com>
> wrote:
>>
>> Hi all,
>>
>> I'm running galaxy-central as my development server, and noticed what
>> to me is a regression with repeat parameters,
>>
>> e.g.
>> https://github.com/peterjc/pico_galaxy/blob/master/tools/clc_assembly_cell/clc_mapper.xml
>>
>> Read group:
>> [+ insert read group]
>> * 1: Read Group
>> * 2: Read Group
>>
>> which on clicking becomes:
>>
>> Read group:
>> [+ insert read group]
>> * 2: Read Group
>> * 1: Read Group
>>
>> This to me is upside down, the current behaviour on galaxy-dist is
>> more natural (and also pluralises the group heading):
>>
>> Read Groups
>> * Read Group 1
>> [Add new Read Group]
>>
>> which on clicking becomes:
>>
>> Read Groups
>> * Read Group 1
>> * Read Group 2
>> [Add new Read Group]
>>
>> Is this a deliberate change? If so, why?
>>
>> Peter