Hi,

On Sun, 2007-11-04 at 20:37 -0500, Daniel Falk wrote:

Photoshop has a tool that works like the healing brush except that it doesn't require a source region to be specified before using the tool. When there are a lot of quick touch-ups to do, this is very convenient.

Photoshop somehow guesses what it should use as source material and is often accurate. When it's not accurate, users can undo it, and then fall back on the healing brush and manually specify that information.

Since we don't know how this works in detail, there is not much point in suggesting that we add such a feature.

In general I would like to point out that it is unlikely that any of the active core developers will pick up your feature requests. If you can find a developer who is interested in the algorithms and willing to work on this stuff, then we are very willing to give him/her a hand and guide him/her through the GIMP code and to review patches. But without a volunteer, this is likely to be just another feature requests. We already have several hundreds of them.

Sven

Hi,

On Mon, 2007-11-05 at 19:29 -0500, Daniel Falk wrote:

Since we don't know how this works in detail, there is not much point in suggesting that we add such a feature.

I could find a video for anyone interested, but that really wasn't my point. I suggested the feature not simply to ask for someone to copy photoshop in detail, but to solve the same problem that photoshop has managed to solve. Namely, figuring out an effective, efficient, and time-saving way of cleaning up a photo with a lot of marks or a dusty scan.

A video wouldn't help. In order to implement this, one would have to know exactly how "Photoshop somehow guesses what it should use as source material". Of course if someone has solved the problem you outlined above, then we would be happy to help him/her to implement it as a GIMP tool.

That's a shame, but I do understand there is a lot of work to be done on the gimp and only so much expertise to go around. Still, can it be logged as a valid feature request somewhere in the event that someone with the interest in improving the gimp might choose to implement this request?

No. It is pointless to keep an enhancement request for something that doesn't have a known solution. It would even be a waste of developers time since this bug would only make our long list of feature requests even longer.

Sven

Sven Neumann wrote:

Hi,

On Sun, 2007-11-04 at 20:37 -0500, Daniel Falk wrote:

Photoshop has a tool that works like the healing brush except that it doesn't require a source region to be specified before using the tool. When there are a lot of quick touch-ups to do, this is very convenient.

Photoshop somehow guesses what it should use as source material and is often accurate. When it's not accurate, users can undo it, and then fall back on the healing brush and manually specify that information.

Since we don't know how this works in detail, there is not much point in suggesting that we add such a feature.

Could you explain the reasoning behind this? Such feature requests are always a good thing, and listening to them is a sign of a user-centric development team. By "listening" to them I don't mean *implementing* them, but a steady stream of such ideas can be beneficial.

Though I suppose suggesting it on IRC might be more appropriate then on the list. Is that what you meant?

Joe

Joe Eagar wrote:

Sven Neumann wrote:

Hi,

On Sun, 2007-11-04 at 20:37 -0500, Daniel Falk wrote:

Photoshop has a tool that works like the healing brush except that it doesn't require a source region to be specified before using the tool. When there are a lot of quick touch-ups to do, this is very convenient.

Photoshop somehow guesses what it should use as source material and is often accurate. When it's not accurate, users can undo it, and then fall back on the healing brush and manually specify that information.

Since we don't know how this works in detail, there is not much point in suggesting that we add such a feature.

Could you explain the reasoning behind this? Such feature requests are always a good thing, and listening to them is a sign of a user-centric development team. By "listening" to them I don't mean *implementing* them, but a steady stream of such ideas can be beneficial.

Hi Joe

Suggesting a new feature without specifying how the algorithm behind it work is pointless because how could the feature then be implemented? There are way too many other things to use the sparse GIMP developer resources for than to research details of other peoples feature requests.

Note the difference between not listening to users and rejecting incomplete feature requests. We appreciate that you think GIMP is worth spending some on to help improving, but please don't take it personal if some of your suggestions are considered incomplete.

It would be very appreciated if you took the time to research exactly how this algorithm is supposed to work.

Regards, Martin Nordholts

Martin Nordholts wrote:

Hi Joe

Suggesting a new feature without specifying how the algorithm behind it work is pointless because how could the feature then be implemented? There are way too many other things to use the sparse GIMP developer resources for than to research details of other peoples feature requests.

Well people do it to me all the time with blender. . .I sometimes figure it out, and if I don't have time to develop it myself I'll try and tell some other dev how it works. And he also offered to show a video about it. Feature videos are really useful for reverse engineering; I don't understand why Sven said otherwise. You can tell a lot sometimes.

Also it's not as if anyone *has* to devote time to figuring it out. Users will make many, many more requests then there will ever be time to code, much less research. Simply listening to the more popular or useful sounding ones will give devs an impression of what users really want, and even what they *need*. This can help formulate long-term plans, both for the project but also for individual devs who think that way.

Such requests might not always be appropriate for a feature tracker, or even a mailing list (I think IRC is a good place to put forth these ideas). But they shouldn't be rejected out of hand, either. I'm not totally sure the best way to handle this; Blender has a functionality mailing list that kindof serves the purpose of random feature requests, but it doesn't work very well. Like I said, for unlikely or somewhat obscure features it seems to be best if users discuss them on IRC, then if a dev gets interested he can, oh I don't know add it to the feature tracker or something like that. Or if he's like me, he may think about these sorts of a features every once in a while, and in a year or two even implement a few of them.

Note the difference between not listening to users and rejecting incomplete feature requests. We appreciate that you think GIMP is worth spending some on to help improving, but please don't take it personal if some of your suggestions are considered incomplete.

Imho, an incomplete feature request is something like "I want a tool to make healing brush better" or something weird like that. "I want a tool that automatically selects a source region for healing brush" imho gives plenty of information, and if passed along to someone who knows the math behind it might even make sense to them. No one is obligated to research this, or even pay attention to it. All that matters is it sounds like a useful idea proposed by a user. And feature videos can help; I myself pieced 3d ray baking together from watching one modo video.

It would be very appreciated if you took the time to research exactly how this algorithm is supposed to work.

Regards, Martin Nordholts

Well I don't have the time for that. :)

Joe

Hi,

On Sun, 2007-11-11 at 01:41 -0700, Joe Eagar wrote:

Though I suppose suggesting it on IRC might be more appropriate then on the list. Is that what you meant?

No, I only meant that filing enhancement requests for this is a waste of time unless more information can be provided.

Sven