CS4 etc.
If it ain’t broke, don’t upgrade it

Margulis on new Photoshop:

On CS4: Four Suggested Rules, Part I
Posted by: "Dan Margulis" DMargulis@aol.com
Thu Oct 16, 2008 8:01 am (PDT)

While out of the country, I was monitoring traffic from list members about the CS4 release.
I wasn't motivated to join in, for the same reason that I've never commented on the overall
merits of CS3. I wrote or posted extended reviews of each of the seven whole-number
versions between Photoshop 3 and CS2. With no more books on the horizon and the
column winding down, there was no further incentive for me to look at beta versions, and
there are more productive ways to spend my time than writing reviews of software, even
Photoshop.

I still feel the same way. I don't want to speak about Photoshop CS4 as a whole. Like every
other release it has certain nice new features and certain disappointments. Instead, I
would like to focus on one specific issue, with a view to making FUTURE releases better
than they otherwise might be. I do this because in one area in which Photoshop has
historically been better than other software products, it has recently slipped badly. For
reasons of length, my comments are split into two posts.

Up until roughly Photoshop CS2, while some releases were better than others, the overall
updating pattern was satisfactory. That is, the product was always stable, speed and
performance continually improved, and care was taken that it would not unduly difficult
for users to migrate to the new version.

These factors are not to be taken for granted.

*Anybody who suffered through the release of Quark 4, Illustrator 9, or the final
deathrattles of GoLive is grateful for Photoshop's consistent stability.

*I've just "upgraded" to the most recent version of Microsoft Excel, running in native mode
on a powerful Intel Mac. It is somewhat disconcerting to find that performance is worse
than in a five-year-old version running in emulation mode on an old G4. That has never
happened with Photoshop.

*When applications are mature, rather than devote time to developing useful new features,
which among other things require talent and run the risk of being found not to work
properly just when the product is about to ship, the developer resorts to jacking around
with the interface, making useless changes that are then fobbed off as making the product
more efficient. The archetype of this type of "upgrade" was Illustrator 7. There were so
many random, stupid changes in the interface that many users (including me) trashed it in
disgust, not wishing to spend the time to relearn the program. By contrast, a Photoshop 4
user confronted with Photoshop CS would have to spend some time learning functions that
didn't previously exist, but he would have no trouble figuring out to use the ones he was
already familiar with, in spite of a slightly retooled look.

From CS2 afterward, Photoshop has continued to shine in the first two categories. With
the notorious exception of Bridge, which as I understand it is not really a product of the
Photoshop development team, the application itself has remained solid. And every version
has continued to be faster than the one previous. CS4's performance, particularly in image
display, is significantly better than CS3's.

It's that third category that's gone seriously downhill. Instead of making meaningful
improvements in the core product, the Photoshop team has branched out on various
tangents. And it has followed the lead of others in eschewing serious improvements in
favor of renamings of commands that don't need to be renamed, reassignment of
functions that doesn't need to be reassigned, and repackaging of existing capabilities
without adding new features. When I see Photoshop CS4, I think Illustrator 7.

Foresight is not always 20/20. When new capabilities are added, one doesn't always know
how they will be used or what flaws in the concept may be. When something like layers, or
adjustment layers, or Camera Raw, or Bridge is introduced, we cannot expect perfection.
The development team has to be given leeway to improve the presentation based on user
experiences, even if the basic idea doesn't change. Renaming, reworking, or redesigning
something one, two or possibly even three releases after it has been introduced cannot be
faulted.

That a function has stabilized, however, is a powerful indication that there's nothing
grossly wrong with it. If the developers want to add significant new functionality, that's
great. If all they want to do is to rework the same capabilities into a different-looking
package, that's not great. It should be seen as what it is, an attempt to justify the
"upgrade" that, in fact, reduces its value.

I would like to suggest that the following four simple rules should be adopted not just by
the Photoshop team, but by every Adobe product and, for that matter, every other
software company in our field.

1. Any component that has existed under its present name for at least three versions may
not be re-named unless its basic functionality has radically changed.

2. The physical location of any component that has existed unchanged for at least three
versions may not be moved significantly, absent an overwhelming, irrefutable reason.
"Compatibility" with other products is not a valid reason.

3. Any keyboard shortcut that has existed unchanged for at least three versions may not
be changed, absent an overwhelming, irrefutable reason. "Compatibility" with other
products is not a valid reason.

4. Any parts of the interface that have been stable for at least three versions may not be
drastically redesigned unless accompanied by significant new capabilities.

Abiding by these simple concepts should be the mark of the skilled, and client-sensitive,
development team. Failure to abide by them identifies the lazy, the corporate politician,
and the incompetent. We've seen in recent releases some violations of these principles,
but nothing like the torrent of pointless changes found in CS4.

I. COMPULSIVE RENAMING.
To verify what is 15 years old, I opened my version of Photoshop 3 (1994). My favorite
command, then as now, was Image: Adjust>Curves.

Some versions later, I forget which, some rocket scientist decided to change Adjust to
AdjustMENT. Why this was thought to be any improvement is unclear. The added value to
the user was nil. It was not particularly inconvenient, either, because most people
understood that Adjust and Adjustment were the same thing. Not all, mind you. I got
several questions from unsophisticated users who were trying to follow early editions of PP
and couldn't find curves. Plus, lots of books and documentation had to be rewritten, and
most Actions or scripts that referenced the old name would not work when it was
changed.

In CS, I think, some other brain surgeon decided the term needed to be AdjustmentS.
Same derisive comments apply. Why would anyone whether the thing is called Adjust or
Adjustment or Adjustments? Just pick one, stick with it, and don't hassle us with pointless
name changes. Many of these Adjustments or whatever they are called were very fine
commands in 1992 but quite pitiful by the standards of what could be done with them
today. If Adobe is so overstaffed that it can assign people to make changes to this name
and all the documentation that goes with it, why can't they assign them to upgrade the
Hue/Saturation interface of Camera Raw (which is better than Photoshop's H/S) and add it
as an Adjust, or Adjustment, or Adjustments?

If this sort of silliness happens once in a while, nobody is going to get too bothered. But
here's the rundown on CS4:

*After six years of being called "Shadow/Highlight", it was deemed necessary to change
the name to "Shadows/Highlights". As usual, no benefit whatever to the user, but make-
work for those preparing documentation or needing to update their Actions. As with
Adjust/Adjustment/Adjustments, it suggests a question. If Adobe has so many people
available with nothing better to do than rename stuff and rewrite documentation, why can't
they let them program (Shadow/Shadows)/(Highlight/Highlights) to work on an
adjustment layer?

*People have had to rotate images since the beginning of time. The various subcommands
that accomplish this are, in Photoshop 3, given the sensible name Rotate. Some versions
later, it was felt necessary to rename it to Rotate Canvas. In Photoshop CS4 it is now Image
Rotation. Furthermore, its location has been significantly changed, in violation of Rule #2
above. No benefit whatsoever to the user. And, unlike Shadows/Highlights, this one has
the potential for confusion. Don't think so? Then consider the next one.

*Photoshop CS4 introduces a new correction command, Image: Auto Tone. Being
interested in color correction myself, I tried it out on a random image, and the result was
impressive. I immediately assembled a suite of test images to find out exactly what it did.
The result? Auto Tone is the command that has been known for 20 years as Auto Levels--
but it isn't even in the same menu location that Auto Levels has been for 20 years.

Auto Levels is referenced and recommended in every textbook, including mine. Yes, it's
primitive, but it has value. What's wrong with leaving it with the name and location it's had
since since before some Photoshop professionals were born? Other than Karnak the
Magnificent, what user, looking for Auto Levels (under Image:
Adjust/Adjustment/Adjustments) will be able to divine that he should be looking
somewhere else for a command called Auto Tone?

If Adobe has so many people with a lot of time on their hands, why can't they just leave
the name and location alone, and instead give Auto Levels Luminosity and Color sliders,
which would greatly increase the utility of the command?

II. HIDE AND GO SEEK.
Photoshop users should not be forced to hunt for familiar routines. The moving, let alone
the renaming, of Auto Levels, should not have been permitted, and would not have been
permitted by my suggested rules.

We first saw this pattern in CS2. The Assign Profile and Convert to Profile commands, after
almost a decade under Image: Mode, wer abruptly moved to the Edit: menu bar. No added
value to the user. It takes an unreasonably long time for users to adjust to such a move. If
Adobe has so many idle personnel that it can assign them to make this move and deal with
the documentation, then it certainly should have enough to add the profile editing that
has been so obviously needed in Photoshop since 1998.

It gets worse in CS4. The addition of Auto Tone/Auto Levels to the Image: menu forces
down certain other items three steps down, which is not unduly difficult to adjust with--
except that the rest of the menu is randomly scrambled. Image: Rotate Image is
gratuitously renamed, but it's also moved up, above items that it previously fell below. So
is Image Size. Apply Image, which is one of the most commonly used commands, goes
almost all the way to the bottom of a lengthy menu, having spent the last 20 years near
the top. This is *not* easy to adjust to. If Adobe has such a staff surplus that it can assign
people to make such a move, it can certainly fix Apply Image so that it works properly in
LAB.

Shuffling long-established commands has less than no benefit to the user. Any
hypothetical gain in productivity in the future would be outweighed a thousandfold by the
loss of productivity in adjusting to the change. Apply Image has been found above Rotate
and Image Size for 15+ years. Since there is no change in functionality, the move would be
prohibited by my suggested rules.

III. THE UNKINDEST SHORTCUT OF ALL.
For 15+ years, the keyboard shortcut to access either individual channels, or individual
channel curves if a dialog is open, has been Command-1 for the first channel (red, L, or
cyan), Command-2 for the second, (green, A, or magenta) etc. These are likely the most
commonly used shortcuts by retouchers. Being aware that the Photoshop team
contemplated changing them, at Photoshop World I attempted to count how many times I
used these shortcuts in front of an audience during my three presentations. The answer:
about a thousand times. I'd estimate that in my work, I use them upwards of 250,000
times per year. That's around 5 million executions of these keystrokes over my career.

This change is galling for several reasons, all of which suggest that the person(s)
implementing them have never made serious use of the program. First, the bland
suggestion that we should just "relearn" the new keystrokes after performing the old ones
several million times. This is roughly like reversing the left- and right-click buttons on the
mouse and suggesting that users should "just relearn".

Second, the "new" keystrokes are not executable by anyone with hands smaller than those
of a gorilla. For the last 15 years, the single most common keyboard shortcut for CMYK
retouchers has been Command-4, to show the critical black channel. This has now been
changed to Command-6. It is physically painful to twist the wrist into this position.
Execute this shortcut tens of times per day--as retouchers commonly do--and carpal
tunnel syndrome would be the inevitable result. Even RGB users now have to use
Command-6s and Command-7s when alpha channels are in use.

Third, for more than 15 years the shortcut to access the first channel in Photoshop proper
has been the same as for the curves dialog either within or without an adjustment layer. In
CS4, insanely, they are now three separate shortcuts.

Fourth, there is no indication of an appreciation of how important these shortcuts are. All
the silly stuff discussed in Section I and II is just that--silly stuff. Annoying and time-
consuming to have to adjust to, and irritating in the sense that Adobe chooses to work on
cosmetics when there is so much substance that *could* be improved upon.

It can, however, be lived with. The channel shortcuts are different. They are too critical to
production to lose. Shortcuts that are physically too difficult to execute are the same as no
shortcuts at all. Losing these shortcuts is therefore a dealbreaker. It would be nearly
impossible for CS4 to include as many new features as would be necessary to compensate
for the loss. I don't know whether they could be changed through the Edit: Keyboard
Shortcuts menu, I suspect not. I have heard rumors that in the shipping release there may
be a mechanism whereby users can restore the traditional shortcuts by means of a plugin.
If so, that would be welcome and would likely change some peoples' minds on whether to
purchase the update.

The fact remains, however, that Adobe was irresponsible even to consider such a change.
There can NEVER be a justification for removing a feature as important as this one, unless
it is so glaringly obvious that no reasonable person could disagree.

The stated reason for changing these shortcuts falls far short. An Adobe evangelist writes
as follows: "And while we're on the topic of keyboard shortcuts, we know that changing
them can be painful, but certain shortcuts 'had' to change in PSCS4. Some changes were
made to bring Photoshop in alignment with the other creative suite applications, for
example, Cmd (Mac / Ctrl (Win) + 1 sets the zoom level to 100% brings PSCS4 into
alignment with Illustrator, InDesign, and Flash)."

The above sentence is hard to parse because of the presence of the quote marks around
"had". The writer may have been expressing disgust with the change. If so, I fully agree. If
the sentence is read without the quote marks, it is an insult to every Photoshop user.
Adobe didn't *have* to do anything, and shouldn't have. Unifying command structure
across Adobe applications regardless of how "painful" it is may seem like a compelling
argument to some beancounter in the marketing department, but it is not relevant to the
existing user base. Using it as a justification for damaging changes in longstanding
features is not acceptable.

Remember, I am only suggesting placing off limits things that have been unchanged for
more than five years--and even then, only if the change is for change's sake, without
affecting capabilities. The reasoning is this:

*If Adobe Pamplemousse has been doing things one way for more than five years and
Photoshop has not, then if somebody really feels the need, Photoshop can be changed
AFAIC.

*If Photoshop has been doing things one way for more than five years and Adobe Oxbelch
has not, then, if the move is felt to be absolutely essential, Adobe Oxbelch should be
changed to emulate Photoshop and not vice versa.

*If Photoshop has been doing things one way for more than five years and Adobe
Snickerdoodle has been doing things a different way for more than five years, then what
dimwit thinks that there's a problem that needs fixing?

If somebody truly feels that every feature needs to be equalized across applications, then
the appropriate action is to create two sets: one to be called Adobe Standard, and one
Traditional, or whatever. If options can be added to InDesign for the benefit of those
familiar with Quark, certainly they can be added to Photoshop for the benefit of those
familiar with Photoshop.

In the second half, a discussion of the remake of the adjustment layers interface.