To comment on the following update, log in, then open the issue: http://www.openoffice.org/issues/show_bug.cgi?id=4219
------- Additional comments from [EMAIL PROTECTED] Wed Jan 3 06:38:50 -0800 2007 ------- AW: I can try. AW: One definite statement from theoretic informatic: Each enlargement will raise error probability quadratically. Let's see: We have signed 32bit integer, one bit for sign, 31bit for value -> 2^31 -> 2147483648. We use 1/100th mm internally in DrawingLayer, togehter with the 3x surrounding page area and a maximum of 120 x 120 cm, for now 120 * 10 * 100 -> +/-120000 are used as values. You need 17 bit to display this (2^17 -> 131072). So we have 31 - 17 -> 14bit calculation space left. This means if there is a calculation step involved which exceeds a multiple of 2^14 (16384), the result will be wrong (!) Think about scalings where You first multiply and then divide using integers or similar stuff... Now let's talk about suggested values: Doubling from 120x120 to 240x240 reduces the reserve by 1 bit and thus doubles the risk of error. Quadrupeling from 120x120 to 480x480 quadrupels the risk. So, when using e.g. 1000x1000 cm instead (and seeing it as 1024 for simplicity) we will need 1000 * 10 * 100 -> 1000000 -> 2^20 (1048576) bits instead of 2^14, so the risk would raise by the factor 2^6 (64). If You ask me, i see two possibilities: (a) Do nothing until we are on double precision arithmetics in 3.0 (b) Remove all bounds and welcome a whole bunch of errors I see no sense in choosing a 'bigger but not too dangerous' bound. Changes will lead to errors, it will just differ in the amount when someone uses that feature. ATM we protect the user in some way, but also not safe and sufficient. So i want to hear Your commments. Keep the limit or remove it completely? --------------------------------------------------------------------- Please do not reply to this automatically generated notification from Issue Tracker. Please log onto the website and enter your comments. http://qa.openoffice.org/issue_handling/project_issues.html#notification --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
