Re: [NEWS] Sawfish has tabbing support (finally

[Teika Kazura <teika lavabit com>]

Hi. 

> Sawfish does now ship tabbing support! 

Congratulations! Let me take this opportunity to express my gratitude
to you, Chris. Despite of your hard work, we often take it as
granted. Your contribution is incomparable (save JSH :).

But let us be professional here, stopping for a moment.

In short, it should be at least marked as 'experimental', maybe be
postponed. Let me explain.

It's clear that tab is still in its cradle. (Sorry, I have not yet
tried tab, but it's still obvious.) It will go under changes, and it
is more than likely that they are incompatible. 

To quote myself, a release is not for the developer or fan's
community, but for the whole world. In addition, it is a window
manager, a user interface. For most people, it is indispensable. But
suppose, one day incompatible changes happen, and they easily require
them hours to adjust. Many people can't forgive it. (Let me quote
myself again, "Imagine, 'What?! ... Sawfish sucks!!! I need M*tacity
or whatever!'") 

How can the goal be reached? For one thing, I think it's reasonable to
demand ourselves that the release should wait for the good documents
of the tab. 

It's for users, of course, but also for developers. Writing process of
a document is a very good testing stone of the understanding and the
quality. Also it gives more power to the coders, since the act
of explanation improves their own understandings. 

If any effort of writing doc results in messy ones, then the design
was wrong. Document includes how to implement tab-capable theme, too. 

Opensource projects too often fail because of the lack or the
insufficiency of the documents. (Even commercial products suck by
idiots' writing.) I remember I often got frustrated in this point
using sawfish/librep. (Ack gives me a partial remedy of it, but it has
nothing to do with rep/sawfish ;)

I'd emphasize the importance of the document. It'd be a bad behavior
to make it as a pretext to indicate tab as 'experimental' (and saying,
'don't worry, it's an RC anyway!'), a pretext to neglect the doc
preparation. Once been idle, the idleness sticks.

If tips are available, it's charming. I liked Rodrigo Amestica's
suggestion on term's names.

# To be more precise, 'experimental: designs may be changed in future
# release' is the appropriate warning.

Techno fans are eager to try new things. They know well and are full
of ideas, so releasing an experimental feature can be a good
act. We'd better devise a way to listen to their voice.

Another way is to plan extensions and improvements, you can't foresee
all though, and to actually implement some. If they fit smoothly to
the existing code, then it indirectly proves that the design is
good. If ad-hoc crooky coding is required, it means 're-design
it. period.' (How cruel! but it's true.) 

Again a general reminder: be a bit patient, first write out your
intent before you code. See what was done and what needed change.
Write out the final design.

On Thu, 22 Jan 2009 23:18:45 +0100, GSR wrote:

> The current ideas seem to be that it is slightly complex, current
> method is a bit hackish and last coder is highly busy. 

To tell you what, before Chris' message appear, I was about to post
something similar to what you read now, and it contained one more
point: if GSR's observation is good (I believe in GSR's words
quality.), then implementing tabs from scratch can be a choice. It
must be not so diffcult, after having one. Let us be confident, we've
already got one work, so anyway we're close to the first goal. Maybe I
have to work on tab, leaving toddler wiki.

> I keep on thinking about it, *but did not get much feedback from 
> last posts.* 

*'s are put by me. Yeah, let's discuss! ... if you have time ...

Conclusion again: a release should feature enough tested tab
implementation. 

I'm all ears. Any comments are welcome.

# Status: I have not yet tested animator. I think I can post a cleanup
# patch after Ian Zimmerman's iconify patch tomorrow.

Sincerely,
Teika kazura