Subject: Re: Upcoming releases
From: Paul Rohr (paul@abisource.com)
Date: Mon Mar 12 2001 - 13:45:16 CST
At 01:47 PM 3/12/01 -0500, Dom Lachowicz wrote:
>I'm working on Normal, Web, and Print Layout views right now. Who knows if
>they will be fully-functional by the next release, but they'll be hiding
>behind some DEBUG statements until then.
Cool. Sounds like the right way to handle it.
>>If you skim the s_TellNotImplemented() messages, you'll see that we have a
>>total of five (5) chunks of the UI which still need some significant work:
>>
>> - Tabs dialog
>> - View Headers and Footers
>> - Lists dialog
>> - Border and shading dialog
>> - Styles dialog
>
>I intentionally removed the Border and shading dialog from non-debug builds.
>It's there if someone wants to do it, but I don't consider it 1.0 critical
OK. I just wanted to make sure that anyone who wanted to defend (and
presumably implement) that feature knew that it was on the chopping block.
Speak now or forever hold your peace.
>and probably more work than it is worth.
What makes it so bad? Getting a complex XP preview implemented for the
dialog, or acually adding the drawing logic to the formatter? If it's the
latter, which is worse -- weird line styles or shading the backgrounds?
>The rest *need* to get done
>(reasonably well) before 0.9.0. We can still fix bugs and tie up loose ends
>and such.
Exactly.
>>Of these, borders and shading and tabs both seem to have fallen off the
>>radar horizon. Is anyone interested in either tackling these or writing up
>>the necessary POWs? If not, then we should make an explicit decision that
>>it's OK to ship 0.9 and 1.0 without them.
>
>See my above comment.
Yep. Without someone to implement or POW borders & shading, it's gonna get
cut.
>I can wholly sympathize with this opinion. I'd like to have as many features
>that "Just Work" as possible. If I (or anyone else for that matter) is to
>agree with your above statement, Paul (as I think that we should), we need
>to lay out exactly what we consider "Just Working" for each feature in
>question and what the acceptable +/- for each is.
Exactly. Since we all know how 100 million people define "Just Works" for
any given feature, it shouldn't be too hard to have these discussions. I'll
assert that *any* divergence from the nitty-gritty details of how similar
features are implemented in competing products should be made explicit so we
can discuss the merits of alternative approaches.
I think the place we run into trouble is when we don't use existing
implementations as a reference point when discussing designs. There's a lot
of prior knowledge we can take advantage of if we don't always set out to
reinvent the wheel from scratch.
This certainly does *not* mean that we have to replicate the mistakes of
other implementations (shudder), but we do need to understand the strengths
of those implementations before we can make an intelligent decision about
whether an alternative is really better.
In an ideal world, whoever takes the lead on implementing a new feature
would:
- examine the competition in detail,
- figure out what it would take to replicate that functionality,
- make reasonable implementation choices, and
- explain the divergences, if any, and how to resolve them.
It's very difficult to spec in functionality afterward when it got
deliberately (or inadvertently) omitted from the original design.
>>Specifically, I'm wary of sneaking Word export into 0.9 unless we expect a
>>very long pre-1.0 release cycle to iron out any and all glitches that could
>>surface. People's expectations that they can round-trip their Word
>>documents to & from AbiWord without *any* loss of data or functionality
>>will
>>be very, very high.
>
>I am of the same opinion. I am going to continue to work on wv on its
>separate release schedule, as always. I will always be syncing Abi's version
>with the official repository. I think that over the next few days I will
>update our exporter so that once my wv routines work 100% correctly, we will
>be able to export to DOC entirely for free. The wvExporter API is almost
>frozen right now. It's simple, but it's a first cut and I'm relatively sure
>that we can do a lot with it. Until then, the exporter will stay hidden
>behind a DEBUG statement. My progress has been hindered since no one else
>has expressed an interest in helping out ;-(
That's what I'd assumed. I'm really eager to start playing with the stuff
you're doing, and I'd love to see more folks helping you. I just wanted to
make sure we didn't inadvertently yoke together that work and 1.0 without
more serious consideration.
>>Thanks to several years work on wv, we already have excellent Word import
>>support, but there are still plenty of potential gotchas that'll really
>>start showing up as soon as we start round-tripping documents. For
>>example,
>>I can easily envision problems like:
>>
>> - lossy styles
>> - mismatched default templates
>> - incompatible fields and/or lists
>> - etc.
>>
>>and I haven't even tried to export anything yet! (This isn't a slam at Dom
>>or anyone else -- there's just a ton of work to be done here.)
>
>Several tons of work, actually. We will most certainly be lossy for certain
>operations.
That's what I was afraid of. This is totally acceptable for developer
versions and debug releases, but the bar for inclusion in 1.0 should
probably be higher.
>>I know that Word export is an ultra-cool feature to add to the bullet list
>>of features we support, but I don't want our marketing to get ahead of the
>>reality. This is a lot of work, and at the moment I suspect it may be more
>>prudent to ship 1.0 without the feature at all.
>
>We'll see about this. It really depends on how much work I get done and how
>satisfied we are with the results. Can we just agree to say that we'll
>re-evalute this a little further down the road?
Absolutely. By all means, keep on hacking and don't stop!!! :-)
Paul
This archive was generated by hypermail 2b25 : Mon Mar 12 2001 - 15:08:23 CST