Cannot set indent on a paragraph to exact values

[VincentBevort]

Hi,

I am trying to set in indent on the body paragraph to 2 cm. (BTW. this goes for all paragraphs)
After clicking the OK button the paragraph shifts in the style editor to what I assume is the correct indent.
To verify the setting I open the style again and see that the value in now 1.88 cm instead of 2.00
not changing anything but clicking OK shifts the paragraph to the left!
Opening the paragraph again I now see a value of 1.78 cm and so on...

Doing the same but now with Units set to points.
I set it to 55 (=2.00 cm according to the paragraph popup).
After opening the units are set back to CM and the value has changes as described above.

After some additional testing this behavior exists on every input box in that Paragraph popup

How can I make the style to remember the setting as I set them?

[Simon Dismore]

I open the style again and see that the value in now 1.88 cm instead of 2.00
not changing anything but clicking OK shifts the paragraph to the left!
Opening the paragraph again I now see a value of 1.78 cm and so on...

I think this is a bug. I could reproduce it in build 3922 but not in build 3790. I found I could prevent the problem by changing View > Program Options > Editor > 'Ruler units' to points or pixels.

[VincentBevort]

Thanks Simon,
Work around works!

[Tim Green]

This phenomenon is caused by the fact that HM uses pixels internally for all dimensions. When you use other units you will often see rounding adjustments in the values that you originally entered when you return to the settings. The only way to get really precise entries is to use pixels is your native units.

[Simon Dismore]

This phenomenon is caused by the fact that HM uses pixels internally for all dimensions. When you use other units you will often see rounding adjustments in the values that you originally entered when you return to the settings. The only way to get really precise entries is to use pixels is your native units.

I don't think an delta of nearly 10 pixels each time is down to rounding adjustment, particularly as it didn't happen in earlier builds. If it looks like a bug, walks like a bug and quacks like a bug...

If there are cumulative rounding errors in H&M, there's no excuse for them. If you define the precision you want it's fairly easy to program conversions so they to guarantee predictable rounding in both directions, so ConvertTo( ConvertFrom( x )) returns x for all values that are integer multiples of the precision. I remember writing this stuff years ago and still have the code somewhere around.

The trick is to store measurements using a high precision intermediate unit, and then do conversion via the intermediate. In the days of VBX components twips (twentieths of a point) were the intermediate, and you could do reversible rounding to 0.001 inches or 0.02mm. That precision meant you couldn't store 0.99mm, because it isn't a multiple of 0.02mm. If you did, it would come back as 1.00mm. But it would always have that value no matter how often you converted it into inches or points or whatever. For Open Office XML Microsoft introduced the English Metric Unit (EMU, 1/914,400 of an inch) which gives predictable results down to 1 μm.

[Martin Wynne]

And another quacking bug.

"centimeters" should be "centimetres".

Martin.

[Simon Dismore]

"centimeters" should be "centimetres".

Not according to my school textbooks...

Textbook.jpg

[Martin Wynne]

Hi Simon,

You appear to have an American text book. In which case all bets are off.

See: https://en.wikipedia.org/wiki/Metre

"Metre is the standard spelling of the metric unit for length in all English-speaking nations except the USA."

In the real world on this side of the pond, a meter looks like this:



Martin.

[Simon Dismore]

You appear to have an American text book. In which case all bets are off.

Actually that's the Nuffield Foundation 2nd Ed Book of Data, published by Longman/Penguin for Chemistry O level. But to be honest I've been using American spelling in technical writing for so long it's become a habit.

Whether we use metres, Angstroms, rods, poles or perches, I'd like to hear Alex's opinion of this "rounding adjustment"

[Tim Green]

I don't think an delta of nearly 10 pixels each time is down to rounding adjustment

I didn't notice the amount of the discrepancy; this does sound a little strange. We're going to look into it. (With known issues one develops an unfortunate tendency to skip the details because they're often irrelevant... )

"centimeters" should be "centimetres".

We've been using US spelling as a standard ever since the first release of Help and Manual.

[Simon Dismore]

We've been using US spelling as a standard ever since the first release of Help and Manual.

Quite right IMO. Sell lots of H&M licenses ⇒ keep improving the product. There's no right or wrong anyway. If you google centimeter site:ac.uk and centimetre site:ac.uk it's the US spelling that predominates at British universities.

[Martin Wynne]

We've been using US spelling as a standard ever since the first release of Help and Manual.

Truly an age of civilisation is drawing to a close. Happy Christmas.

Martin.

[Alexander Halser]

We fixed the problem with build 3925:
http://www.helpandmanual.com/downloads.html?upd

[VincentBevort]

Thanks,


Works great