18-DEC-1997 11:22:59.79
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5.10.5b. Result of MEASURE, 3D graphics.
Dear Users:
1. Now FFS["MEASURE"] returns the center of mass and beam matrix as a list:
{center-of-mass, beam-matrix} ,
where center-of-mass is a vector containing the center of mass of
{x, px, y, py, z, delta}, and beam-matrix is is a 6 by 6 symmetric matrix
having the beam matrix.
2. A very preliminary verion of 3D graphics has been installed. It works only
with Tkinter and CanvasDrawer. Now only 2 functions works:
a) ListPlot3D[list, options, ...] where list is a list of data
{{x1, y1, z1}, ...} .
Currently available options are PlotRange, PlotRegion, Box, EyePoint.
b) ListBirdsEyePlot[{{x1, ..., xn},
{{z1, {y11, ..., y1n}}, ..., {zm, {yn1, ..., ymn}}}, options]
Currently available options are PlotRange, PlotRegion, Box, EyePoint,
Lighting, and LightSources. LightSources is a list of list
{{r, phi, theta},{r, g, b}} to give the distance angles, and rgb color of
the light sources. It is active when Lighting is True (default). Multiple
light sources are possible.
K. Oide
8-DEC-1997 12:00:52.44
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5.10b. Definition by ?, etc.
Dear Users:
1. Now ?expression works same as Definition[expression], unless expression is
one of names of FFS flags. Expressions like as ?CELL returns the value of FFS
flag CELL as before.
2. An internal logic for Table has been wrong when Table picks up a value of a
symbol which is assigned in the Table itself. The simplest was something like
Table[s={i},{i,2}] .
This bug caused inconsistency in the memory usage of the interpreter. This
bug is very old probably since the beginning of the SADScript interpreter. It
is interesting that such a very primitive bug has survived for a long time.
K. Oide
4-DEC-1997 21:12:30.99
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5.9.2b. A bug.
Dear Users:
1. A bug was found in output of error message during Get. Such an error
message has been suppressed by this bug. This bug also has caused a
destruction of memory (internal free area map), then very unpredictable
things happened. By fixing this bug, things look OK so far, when an
error occurs in Get, but probably we have to watch it for the time being.
I also added a routine in MemoryCheck[] to see the consistency of the
free area to detect this kind of problem.
K. Oide
3-DEC-1997 13:38:27.26
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5.9b. StringToStream, etc.
Dear Users:
1. A new function StringToStream has been added. It is same as
Mathematica's, and opens an input stream from character-string instead
of a file.
Usage: fn = StringToStream[string];
Read[fn, Word];
Close[fn];
With this modification, the internal logic for Open/Read is greatly
updated.
2. New options NameFilter and Initialize have been introduced in
OpticsPlot. NameFilter -> function specifies a function to replace the
name of components to be displayed with function[name]. Names are now
diplayed in semi-vertical way in CanvasDrawer.
Initialize works only for CanvasDrawer. Initialize ->
{{xmin,xmax},{ymin,ymax}}, where coordinates are in the
relative window coordinates ({{0,1},{0,1}} is the graph area), erases
items enclosed in the specified rectangle. Initialize -> Frame erases
upper part of the canvas. Initialize -> Graph erases the graph area.
3. A bug when a shell command is executed by OpenRead with "!" mark is fixed.
This bug stops entire program when OpenRead is done more than 2048 times.
4. DISPLAY now displays the name of the first component of an element with
".1" attached, when there are 2 or more components of the element exist in the
beam line.
K. Oide
27-NOV-1997 02:23:26.53
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5.8b (cnt'd) PRINTGetMAIN, comment by (
Dear Users:
1. There are a few more updates on V1.0.5.8b which was not mentioned in the
previous mail. One is a new function GetMAIN. GetMAIN[file] reads SAD MAIN
commands such as lattice definitions from file. GetMAIN ignores FFS, EMIT,
TRACK commands in the file. The parameters of components in the current line
which FFS is using is not updated until RESET is issued.
2. Now any charaters enclosed by (* and *) are recognized as a comment. This
is the Mathematica style. Comment can be laid through many lines.
K. Oide
25-NOV-1997 19:09:42.16
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5.8b. StringTrim, Fit, etc.
Dear Users:
1. Bugs in Fit have been cured. This bug caused trouble when initial values
of the fitting parameters are very close to the final result.
2. A new function StringTrim was introduced. It removes leading and trailing
spaces and tabs from a character string.
3. Expression like {a, b, c, ..} = {d, e, f, ..} now works when a, b, c, ..
have upvalue definitions (a = ..)^ = .. .
K. Oide
16-NOV-1997 09:32:37.31
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5.7.2b. Bug in matching.
Dear Users:
1. A bug was found in the optics matching, when ROTATE was a variable.
This bug was introduced when the flag CANON became obsolete, around
6/5 this year. The speed of matching was very slow or did not match
at all. The result of optics calculation itself was OK.
K. Oide
16-NOV-1997 07:52:21.13
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5.7.1b. Bug in off-momentum matching.
Dear Users:
1. Since around 10/30 this year, the off-momentum matching has been wrong,
using a wrong momentum offset for off-momentum optics.
2. A function Literal has been introduced since the last version. It is same
as Mathematica's to block evaluation for pattern matching, esp. for lhs of
Rule or RuleDelayed.
K. Oide
15-NOV-1997 04:48:27.57
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5.7b. Bug in Times, etc.
Dear Users:
1. A bug is found in an expression like as
a * (b + c + d + ...)
when a is a number and the number of terms for Plus is 3 or more. It wrongly
calculated a * (b + c) and components after third were omitted. The time when
this bug came is not easy to track, but it is not very new, at least 6 months
before. Also expressions like (a ^ b ^ c ...) ^ d were wrong when the terms
in the parenthesis was 3 or more.
2. Break[] did not work properly in While, leaving a wrong message. This
bug was created around 4/26/1997.
3. The rouine to refer a function which contains definitions without patterns
is updated to improve the speed when the number of definitions is large.
K. Oide
6-NOV-1997 10:42:58.78
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5.6b. TagSet, etc.
Dear Users:
1. A new operator TagSet has been introduced. There are threee variations in
its usage:
symb/: lhs = rhs
symb/: lhs := rhs
symb/: lhs =.
These define/undefine lhs, assciated only with the symbol symb. For instance,
c/: a[b, c, d] = 3;
defines the upvalue for c, leaving a, b, and d untouched. TagSet is
effectively equivalent to Mathmatica's, but SAD does not have TagSetDelayed and
TagUnset as separate operators.
2. There has been a problem in EMIT when RFSW is OFF and RADCOD is ON, since
this July. The caculated COD with such setting of flags has been wrong. Now
the COD is calculated as if there is an acceleration at the beginning of line
to compensate the energy loss.
3. Extract[a, {i, j, ...}] was wrongly coded so that it did not evaluate after
extraction. Now it is corrected to evaluate.
K. Oide
30-OCT-1997 15:16:33.32
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5.5.b. DDPI as a matching variable.
Dear Users:
1. Now the initial momentum offset can be a matching variable, only for orbits
DX, DPX, DY, DPY.
Usage: FREE DDPI
Now the central momentum DP0 changes as the result of matching if DDPI becomes
a variable. Actually DP0 is equivalent to LINE["DDP",1].
K. Oide
15-OCT-1997 09:36:23.79
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5.4b. Beam sizes by LINE.
Dear Users:
1. The LINE function now returns beam-matrix element by "SIGab" keyword, where
a and b are one of X, PX, Y, PY, Z, DP. If b is omitted, it returns
sqrt of SIGaa, i.e., the usual beam size. If both a and b are omitted, it
returns 6 by 6 beam matrix. The second argument of LINE can have a fractional
part which means the fractional position (same as for "S" and "GAMMA").
The beam size is caluculated with twiss parameters given by CALC and numbers
in EMITX, EMITY, and DP. Thus it is on the 4 by 5 basis. All quantities in
z-derection are not calculated and meaningless at this time.
2. OpticsPlot now accepts function names "SIGab" as described above to denote the
beam sizes.
K. Oide
7-OCT-1997 10:47:03.13
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5b. STACKSIZ.
Dear Users:
1. The evaluation stack of the interpreter is now allocated dynamically.
It reduces the size of the load module by about 3Mbytes, and also its size
becomes configurable depending on applications.
STACKSIZ = nnn;
sets the size of the stack to nnn. This must be issued before the first FFS
session, othewise does nothing. If you have a problem on the memory size
during execution of SADScript, try this. The default is 200000.
2. A new attribute Constant is defined. If a symbols has Constant attribute,
it is evaluated at the input time immediately. Symbols such as True, False,
Pi, etc., and also all built-in functions are declared Constant. Therefore an
expression like
Hold[Cos[Pi]]
always becomes Hold[-1].
K. Oide
7-OCT-1997 10:46:45.89
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.5b. STACKSIZ.
Dear Users:
1. The evaluation stack of the interpreter is now allocated dynamically.
It reduces the size of the load module by about 3Mbytes, and also its size
becomes configurable depending on applications.
STACKSIZ = nnn;
sets the size of the stack to nnn. This must be issued before the first FFS
session, othewise does nothing. If you have a problem on the memory size
during execution of SADScript, try this. The default is 200000.
2. A new attribute Constant is defined. If a symbols has Constant attribute,
it is evaluated at the input time immediately. Symbols such as True, False,
Pi, etc., and also all built-in functions are declared Constant. Therefore an
expression like
Hold[Cos[Pi]]
always becomes Hold[-1].
K. Oide
25-SEP-1997 20:28:29.06
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.4b. Message.
Dear Users:
1. The message handling capability of SAD has been enhanced, and now it looks
very similar to Mathematica's. The new operators and functions are:
symbol::tag
symbol::tag = message
MessageName
Message[symbol::tag]
Message[symbol::tag, arg, ...]
$MessageList
MessageList
On[symbol::tag]
Off[symbol::tag]
On[General::newsym]
Check[exp1, exp2, symbol::tag, ...]
These are basically same as Mathematica's. Please look at the Mathematica
Book or SAD/Tkinter manual (in Japanese).
K. Oide
25-SEP-1997 20:28:20.79
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.4b. Message.
Dear Users:
1. The message handling capability of SAD has been enhanced, and now it looks
very similar to Mathematica's. The new operators and functions are:
symbol::tag
symbol::tag = message
MessageName
Message[symbol::tag]
Message[symbol::tag, arg, ...]
$MessageList
MessageList
On[symbol::tag]
Off[symbol::tag]
On[General::newsym]
Check[exp1, exp2, symbol::tag, ...]
These are basically same as Mathematica's. Please look at the Mathematica
Book or SAD/Tkinter manual (in Japanese).
K. Oide
10-SEP-1997 09:52:24.86
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.11b. ColumnPlot.
Dear Users:
1. A new plotting function ColumnPlot has been built to create a bar or column
graph.
Usage: ColumnPlot[data, options, ...]
plots a column plot.
1) If data is a 1D vector, it makes a simple column plot.
2) If data is a 2D matrix, it makes a multiple-column plot.
3) If data is a 3D list, it makes a stacked, multiple-column plot.
Besides options common for all plotting functions, ColumnPlot has its own
options:
Option Value Default Action
---------------------------------------------------------------------------
ColumnOffset 0
30-AUG-1997 19:08:40.37
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.10b. SAD/Tkinter Manual, etc.
Dear Users:
1. A manual of SAD/Tkinter is now ready for download from the SAD homepage:
http://www-acc-theory.kek.jp/SAD/sad.html .
2. A new usage of Part is introduced. If an index is omitted or Null,
a list of elements with all ranges of the omitted one having other indices
as specified. For instance,
a = {{1,2},{3,4}};
then
a[[,2]] returns {2,4}.
Thus a[[,k]] is equivalent to Thread[a][[k]] for a matrix, but runs much faster.
3. Variables in matching are now restricted in the MIN/MAX range, even started
from outside of he range.
4. Logics of scoping of variables are updated.
K. Oide
15-JUL-1997 18:44:41.62
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.9b. In[n]:= as prompt, OpenRead, TemporaryName, .
Dear Users:
1. The style of the prompt for the terminal input is changed to
"In[n]:= ", where n is the line number of the next output. Note that it is
just appearance, and In[n] does not retain the input. Also the handling of
the output is somewhat improved, when the output of the script is a symbol or
an expression causing then "Undefined command or element" message. If you
like the old style of the prompt, say FFSPRMPT.
2. The function OpenRead was so changed that it returns $Failed on error at
opening a file. Also a form like
f = OpenRead["!command"]
returns a file number f which is used to read the output of command.
3. A function TemporaryName[] is introduced to return a unique name of a
temporary file.
4. A function Skip is installed. The usage is same as Mathematica's.
K. Oide
9-JUL-1997 11:42:08.14
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.8.1b. Read[f, {type1,..}].
Dear Users:
1. Now Read can be used as
Read[f, {type1, type2, ...}, option] ,
where type1,.. are the types of objects to be read. This returns a list of
objects read from the file. Currently allowed types are Real, Expression,
Word, and String. options are WordSeparators, ReadNewRecord, and NullWords.
You will also notice a significant improvement of the speed of Read.
K. Oide
8-JUL-1997 10:21:17.83
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.8b. FFS without USE=LINE, etc.
Dear Users:
1. FFS can be started without defining beam line. You can say just "FFS;" to
start it. Thus a SADScript can be executed almost directly.
Example:
FFS;
(FFS commands)
....
ABORT
2. A few improvements have been made for Read, Plot, Fit, etc., but the user
interface is not changed.
3. A new function Vectorize and Vector has been introduced to speed up the
internal processing of a function, but it will be limited to the internal use
for a time being.
4. In DPAERT and DynamicApertureSurvey, the minimum values of X and Y have been
set to zero when the size of the range is zero. This was not a bug, but
deliverately so coded. Though I do not remember the reason, I have removed
this feature.
K. Oide
24-JUN-1997 22:54:58.00
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.7b. Random[n], GaussRandom[n], etc.
Dear Users:
1. Random and GaussRandom now takes arguments to return multiple random numbers
at once.
Random[] gives a uniform random number between 0 and 1.
Random[n] gives a list of n uniform random numbers.
Random[n1, n2, ..] gives a (n1 * n2 * .. ) tensor of uniform random numbers.
and
GaussRandom[] gives a Gaussian random number with average 0,
standard deviation 1, cutoff at GCUT.
GaussRandom[n] gives a list of n Gaussian random numbers.
GaussRandom[n1, n2, ..] gives a (n1 * n2 * .. ) tensor of Gaussian random
numbers.
2. EMIT now returns damping times for convenience.
3. A bug was found in the extended Twiss parameters calculated by EMIT.
The formula on the web page is right, but in the coding a factor 1/(1+a) has
been missed in the x-y terms of the matrix H. Thus the parameters were wrong
in particular when zetas are nonzero.
4. A bug was found in Do with exceptional exit (by such as Return[]).
This caused memory crash. This bug was generated on 4/14/1997, V1.0.2.1b.
K. Oide
22-JUN-1997 21:38:54.01
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.6b. Extract, etc.
Dear Users:
1. Functions Extract, Symbol, and SymbolName have been installed. These are
same as Mathematica 3.0's.
2. Matching by GO has been wrong for weeks, when non-standard variables such as
K0 of BEND are chosen.
K. Oide
22-JUN-1997 12:15:54.33
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.6b. Extract, etc.
Dear Users:
1. Functions Extract, Symbol, and SymbolName have been installed. These are
sam as Mathematica 3.0's.
2. Matching by GO has been wrong for weeks, when non-standard variables such as
K0 of BEND are chosen.
K. Oide
17-JUN-1997 15:08:20.45
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.5.2b. Bug in CAVI.
Dear Users:
1. A bug was found in CAVI in EMIT (when RFSW is OFF), CAL and GO (when RFSW is
off or RING is ON). This bug makes the length of CAVI 1/2!!! This bug was
created after 6/5, V1.0.3b, probably last week. Before that time there had
been no such problem.
K. Oide
16-JUN-1997 21:57:29.81
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.5b. Acceleration in MULT, etc.
Dear Users:
1. MULT now accepts acceleration as CAVI. The parameters are VOLT, PHI, FREQ,
and HARM. This is useful to express a trim winding on an accelrating
structure.
2. The transformation of BEND in optics calculation and EMIT has been changed.
It is now as exact as that of tracking, and no limitation in the size of the
edge angle.
3. A few commands which accepts a number now accepts a Set form. For
instance,
SEED n can be written as SEED = n or seed = n.
Such commands are AXI, BXI, AYI BYI, EXI, EPXI, EYI, EPYI,
R1I, R2I, R3I, R4I, DXI, DPXI, DYI, DPYI, DZI, DDPI, and MAXI.
4. There were a few bugs created last week which is related to the item 1
above. Those now look fixed.
K. Oide
10-JUN-1997 13:20:04.00
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.4b. Limits in FindRoot and Fit.
Dear Users:
1. Limits on variables or fitting parameters now can be give as
FindRoot[eqs, {var1, ini1, {min1, max1}},..]
or
Fit[data, fun, var, {par1, ini1, {min1, max1}}, .. ]
2. A problem with SEED and GaussRandom was solved. The value of GaussRandom
was not uniquely determined by SEED, but now it was solved. This problem was
only for SEED, and there has been no problem in SeedRandom.
K. Oide
9-JUN-1997 12:12:27.36
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.3.3b.exe. FitPlot.
Dear Users:
1. A new function FitPlot has been installed.
Usage: FitPlot[data, fun, var, {par1, ini1}, .., {parn, inin}, option]
makes a plot of data with a fitted function fun, with parameters par1,
parn. Specification of data is same as ListPlot. This is just an
application of Fit, but quite useful. Please try.
K. Oide
8-JUN-1997 17:40:28.54
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.3b. Fit, Out, etc.
Dear Users:
1. A new function Fit has been installed. Unlike Mathmatica's it does
general nonlinear fitting of data.
Usage: Fit[data, expr, var, {par1, ini1}, .. ,{parn, inin}]
performs a nonlinear fitting of data with an expression expr.
data: list of {xi,yi}, {xi,yi,dyi}, or {xi,yi,dxi,dyi}, where dxi and dyi
are the standard deviation of the i-th point.
expr: an exprssion containing var as the x-variable, and
parameters par1,..,parn explicitly. Use Evaluate[fun] if an implicit
expresison is necessary.
var: a symbol to express the x-axis variable.
par: parameter symbol to be varied in the fitting.
ini: initial value of the parameter. It must be specified.
Fit returns the result as a list:
{par1 -> v1, .., parn->vn, ChiSquare -> chisq, GoodnessOfFit -> good,
ConfidenceInterval -> {c1, .., cn}, ConvarianceMatrix -> cov},
where v1,..,vn are the values of the parameters which minimizes chi-square,
chisq is the resulting minimum value of the chi-square, good is a number
given by GammaRegularized[(ndata-npara)/2,chisq/2] to represent the goodness
of the fit, c1, .., cn are the estimated errors in parameters, and cov is
the covariance matrix. A typical criterion of the goodness is (good > 0.1).
The current version of Fit does not handle the error in x.
2. Now all results of evaluation of an input line is stored in Out[n]. This
is just the same mechanism as Mathematica's. Out[n] can be written as %n. A
single % means Out[$Line], i.e., the last output, where $Line is the counter
of the output line number. If one reset $Line to a number smaller than current
value of $Line, the contents of Out between them are cleared.
3. A new operator @ has been introduced. It is same as Mathematica's, i.e.,
f@a is same as f[a].
K. Oide
7-JUN-1997 10:22:27.37
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.2.1b. Bug in f[..][..].
Dear Users:
1. Another bug in an expression f[x,..][y,..] was found. This occurred when
the expression is either in the definition of function, or an argument of Do or
Table in Module, containing the argupent pattern or the step variable in
f[x,..], and also y,.. are all real numbers.
K. Oide
7-JUN-1997 10:22:10.51
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3.2.1b. Bug in f[..][..].
Dear Users:
1. Another bug in an expression f[x,..][y,..] was found. This occurred when
the expression is either in the definition of function, or an argument of Do or
Table in Module, containing the argupent pattern or the step variable in
f[x,..], and also y,.. are all real numbers.
K. Oide
5-JUN-1997 19:59:38.97
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3b (supplement). Erf, Erfc, GammaRegularized, etc.
Dear Users:
1. V1.0.3b has also installed functions Erf, Erfc, GammaRegularized,
GammaRegularizedP, GammaRegularizedQ. These are currently valid for real
arguments. Erf, Erfc, and GammaRegularized are same as Mathematica's.
GammaRegularizedQ is same as GammaRegularized. GammaRegularizedP is equal to
1 - GammaRegularizedQ.
K. Oide
5-JUN-1997 19:16:04.79
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.3b. InitialOrbits, WakeFunction, etc.
Dear Users:
1. The new version 1.0.3b has two major upgrades in the calculation of optics
in FFS. One is InitialOrbits to set a number of arbitrary initial orbits.
Usage: InitialOrbits = { {x1, px1, y1, py1, z1, dp1}, ...};
specifies initial conditions of a number of orbits for the optics
calculation by CALCULATE(CAL) and GO. Those coordinates are offset from
the central orbit. If InitialOrbits are given, the off-momentum matching
and finite-amplitude matching is disabled. InitialOrbits is also necessary
to calculate optics with wake field.
2. The second change is optics calculation with wakefield:
WakeFunction[Longitudinal, comp]={{z1, wl1}, ..., {zn, wln}};
WakeFunction[Transverse, comp]={{z1, wt1}, ..., {zn, wtn}};
specify logitudinal and transverse dipole wake functions at a component comp
(string). Each functions is a list of {z, w} where z is the distance (z>=0)
and w is the value of the wake, in unit of either V/C or V/C/m.
The wake funcitons are applied at the componet comp, giving kicks to each
orbit whose initial conditions are give by InitialOrbits. The sufficient
number of orbits depends on the situation.
WakeFunction is valid only when TRPT and INS, and also either TWAKE or
LWAKE is ON.
Utility functions to use WakeFunction easily are to be supplied soon.
3. New optical functions DZ and DDP are installed in FFS. These represent the
logitudinal position z = -v(t - t0) and the momentum offset dp/p0. These can
be displayed by DISP A.
4. Flags CANON and DIFCOD have become obsolete.
5. SCAN command has become obsolete.
6. A bug in some cases of evaluation of an expression such as f[a,..][b,..] was
fixed.
K. Oide
30-MAY-1997 06:08:09.68
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.2.9.2b.exe. Bugs in Thread, XIX, XIY.
Dear Users:
1. A bug was found in XIX and XIY to make the chromaticity goal half. This
bug looks very old and hard to track.
2. A bug to destroy a memory region is found in Thread. It is generated in
the last weekend.
K. Oide
27-MAY-1997 04:07:40.01
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.2.9b. NullWords in Read.
Dear Users:
1. A new option NullWords is added to Read[f,Word]. When NullWords->True,
as Mathematica, Read[f,Word] returns "" between successive WordSeparators.
K. Oide
17-MAY-1997 03:09:13.74
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.2.7.1b.exe. Bug in K0 of BEND.
Dear Users:
1. Bugs are found in BEND, when ANGLE=0 and K0<>0. The transformation of y
and z was wrong. It was wrong but still symplectic. The discrepancy
becomes significant when K0 is large, and the ordef of the error is in K0^3.
Also another bug was in MARK with OFFSET<>0, and the MARK with OFFSET points
inside a BEND.
K. Oide
14-MAY-1997 11:53:02.03
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.2.7b. FindRoot,AppendTo,PrependTo.
Dear Users:
1. New functions AppendTo, PrependTo are added. These are same as
Mathematica's.
2. A new function FindRoot has beed added as a trial version.
Usage:
FindRoot[a[x,..]==b[x,..], {x, x0}, {y, y0},..., option]
or
FindRoot[{a[x,..]==b[x,..],c[x,..]==d[x,..]..},{x, x0},..,option] .
The number of equations is unlimited. The number of variables is currently
limited less than 1024. Although currently FindRoot is limited to Newton's
method, it might be a littl advanced than Mathematica's, because it
automatically determines the damping factor by a cubic interpolation.
No bounded search on the variable are possible.
There are two options in FindRoot.
MaxIterations maximum number of iterations. (default 50)
AccuracyGoal relative accuracy of the convergence (default 1e-20)
This version of FindRoot is a trial version. Please try it and feedback to
us.
K. Oide
17-MAY-1997 03:08:52.46
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.2.7.1b.exe. Bug in K0 of BEND.
Dear Users:
1. Bugs are found in BEND, when ANGLE=0 and K0<>0. The transformation of y
and z was wrong. It was wrong but still symplectic. The discrepancy
becomes significant when K0 is large, and the ordef of the error is in K0^3.
Also another bug was in MARK with OFFSET<>0, and the MARK with OFFSET points
inside a BEND.
K. Oide
14-MAY-1997 11:52:30.63
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.2.7b. FindRoot,AppendTo,PrependTo.
Dear Users:
1. New functions AppendTo, PrependTo are added. These are same as
Mathematica's.
2. A new function FindRoot has beed added as a trial version.
Usage:
FindRoot[a[x,..]==b[x,..], {x, x0}, {y, y0},..., option]
or
FindRoot[{a[x,..]==b[x,..],c[x,..]==d[x,..]..},{x, x0},..,option] .
The number of equations is unlimited. The number of variables is currently
limited less than 1024. Although currently FindRoot is limited to Newton's
method, it might be a littl advanced than Mathematica's, because it
automatically determines the damping factor by a cubic interpolation.
No bounded search on the variable are possible.
There are two options in FindRoot.
MaxIterations maximum number of iterations. (default 50)
AccuracyGoal relative accuracy of the convergence (default 1e-20)
This version of FindRoot is a trial version. Please try it and feedback to
us.
K. Oide
10-MAY-1997 19:06:30.50 MAIL
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.2.6b.
Dear Users:
1. The upgrade this time is mostly on the improvement of the performance of the
interpreter, esp. in pattern matching, replace, list operations, etc. This
was necessary to make the new Tk interface as fast as possible. Actually it
has become more or less acceptable in speed for regular panels. You may
look at /SAD/share/Packages/Tkinter.n and /SAD/share/Packages/Canvas.n how
Python/Tkinter is embedded in SAD. Some examples are available in
/users/oide/WORK/oldsad/sad/examples.sad or
/users/oide/WORK/oldsad/sad/canvas.sad .
2. A new function Pause is installed. It is same as Mathematica's and accepts
fraction of seconds, down to microsecond. Sleep has become an alias of Pause.
3. A bug in complex numbers, such that I==1 or I>1 returned I, was fixed.
K. Oide
26-APR-1997 11:44:55.63
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.2.4b. OpticsProlog and OpticsEpilog.
Dear Users:
1. New variables OpticsProlog and OpticsEpilog are defined.
OpticsProlog is a variable to assign a user-defined function which is to be
executed everytime before an optics calculation is done in CALCULATE(CAL)
or GO commands. In GO, OpticsProlog is called at the beginning of each
iteration. This function is useful, for instance, for setting parameters
which depends on the result of optics calculation.
OpticsEpilog is a variable to assign a user-defined function which is to be
executed everytime after an optics calculation is done in CALCULATE(CAL)
or GO commands. In GO, OpticsEpilog is called at the end of each
iteration. This function is useful, for instance, for setting parameters
which depends on the result of optics calculation.
K. Oide
21-APR-1997 16:10:03.86
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.2.2b. With, SwitchCases, etc.
Dear Users:
1. New functions With and SwitchCases are introduced.
With[{x, .., y=a, .. }, body]
evaluates body assuming the value of x to be the current definition, or
r.h.s. of Set ot SetDelayed, such as y=a or y:=a. In the latter cases,
it is similar to Mathematica's With. With is useful to define a function
efficiently.
SwitchCases[list, {case1,..}]
returns a list { list1, .. }, where list1 is a list of subexpressions which
match case1, etc. SwitchCases does what are done by Cases and DeleteCases
simultaneously. If the second argument is like {c1, .. , _}, the last of the
returned list contains subexpressions which match none of c1, ...
2. The local variable in Module or Block now allow a form like {a,b,c}={u,v,w}.
This is not allowed in Mathematica 2.2, but convenient.
K. Oide
17-APR-1997 20:52:53.57
From: KEKVAX::YAMAMOTO "Noboru"
To: @USER$DISK:[OIDE]SADUSERS.
CC:
Subj: SADUPDATE: CaOpen etc on DEC machine.
Dear SAD users,
The executable module in /SAD/bin is updated to support EPICS CA access
on DEC machines. version number is still SAD1.0.2.1.1b ( it should be
SAD1.0.2.1.2b).
Noboru
14-APR-1997 09:23:43.07
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.2.1b. Formating, etc.
Dear Users:
1. The routines for formatting output was updated to improve the conversion
efficiency. The change in the output should be small, but not zero.
K. Oide
20-MAR-1997 13:27:03.69
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.1.7.1b. Scoping, etc.
Dear Users:
1. A problem in scoping of symbols was fixed. There was a problem a script
such as
h[a_]:=a;
SetAttributes[h,HoldAll];
f[a_]:=h[Module[{},Do[i,{i,10}]]];
f[1]
is executed, the scoping of symbols in unevaluated arguments of a function (the
symbol i in the example above) was wrong. This fix required change of the
internal logic of scoping, thus any undesirable side effects might be expected.
2. Matching of dispersions with orbit change was updated.
3. A bug in VariableRange in the case of returning a list was fixed.
K. Oide
9-MAR-1997 20:15:58.06
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.1.6b. Factorial, matching, etc.
Dear Users:
1. Matching routines for dispersion are improved, when a variable changes the
orbit during the mathcing, such as K0 of a bend, etc. Before this change,
converging speed was terribly slow.
2. New functions Factorial, Gamma, LogGamma are introduced. They accept
real or complex numbers.
K. Oide
3-MAR-1997 20:00:25.72
From: KEKVAX::KIKUCHIM
To: @USER$DISK4:[OIDE]SADUSERS.
CC:
Subj: sad/cor update: Type option in Steer
Dear users,
Steer[...] has now an option of Type:
Steer[pattern,error,option]
At present the option Type is available.
Example.
Steer["OCT*",,Type->22] ! elements having name of "OCT*" and
! type of 22.
Steer["OCT*",,Type->(LINE["type",#]==22&)] ! same as above.
Steer["Z*",,Type->(LINE["k0",#]>1e-3&)] ! element Z* with K0 greater
! than 1e-3.
Defualt is Type->(LINE["type",#]>0&) .
27-FEB-1997 19:57:11.57
From: KEKVAX::KIKUCHIM
To: @USER$DISK4:[OIDE]SADUSERS.
CC:
Subj: sad/cor update: Weight option in CorrectOrbit
Dear users,
I have changed the Weight option in CorrectOrbit[] as follows.
CorrectOrbit[kind,mon,str,optics,Weight->weight]
1. If weight is a numerical list, it is taken as a weight of
monitors in the minimization.
2. Weight allows a function that maps on to mon[1] like
(Twiss["bx",#]&) .
3. Default is (Twiss[kind,#]&) .
KIKUCHI M.
26-FEB-1997 16:03:37.69
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.1.4.5b. Bug in Emittance[Emittance->False].
Dear Users:
1. A bug was found in Emittance[Emittance->False]. Calculation of transfer
matrix, etc. was wrong when this option is used. Without this option, there
is no problem. This bug was created probably on 1/11/96.
K. Oide
24-FEB-1997 12:13:21.98
From: KEKVAX::YAMAMOTO "Noboru"
To: @USER$DISK:[OIDE]SADUSERS.
CC:
Subj: Operation of SAD computer cluseter resumed.
Dear SAD users,
Operation of SAD computer is resumed. SAD computer is available for
normal use.
Noboru Yamamoto
24-FEB-1997 10:16:22.61
From: KEKVAX::YAMAMOTO "Noboru"
To: @USER$DISK:[OIDE]SADUSERS.
CC:
Subj: SAD computers is not available NOW>
Dear SAD-users,
SAD computer is not availabe now. It will be available
1hour later.
Sorry for incovenience.
Noboru
24-FEB-1997 09:55:07.03
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SADV1.0.1.4.3b. Bug in MARK in SOL.
Dear Users:
1. A bug was found in MARK, MONI, APERT when they are placed
between SOLS, with an off-centered orbit. This bug was
created on 7/12/1996.
K. Oide
19-FEB-1997 18:58:46.27
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.1.4.2b. Bug in FitFunction & parallel.
Dear Users:
1. A bug was found in FitFunction when it is processed in parallel (on acsad
when number of variables is larger than 3). This bug was created on 2/13 in
V1.0.1.2b.
K. Oide
15-FEB-1997 21:49:03.50
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update V1.0.1.4.1b. Bug inHP-UX version.
Dear Users:
1. The new version of parallel processing in optics calculation of FFS did not
work on ahsad, when off-momentum matching was tried. It has been corrected.
Katsunobu
14-FEB-1997 12:24:18.97
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.1.4b. Bug in BEND with ANGLE=0 and K0<>0.
Dear Users:
1. A bug was found in routines for BEND with ANGLE=0, nonzero K0, when
the sum of incoming angle and outgoing angle of the beam is zero or very
close to zero. When it was zero, the vertical tracking, optics and
emittcance calculation was wrong. When it was very close to zero, the
accuracy of those thing might be lost. This bug was created when SAD
was on HITAC, for tracking and emittance. For optics calculation in FFS,
the bug was transferred from emit to FFS on 2/21/1996.
2. Another small bug was created yesterday, in off-momentum matching
with CELL and nonlinear elements.
K. Oide
13-FEB-1997 19:21:17.41
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. V1.0.1.3b. Bug in fraction of component.
Dear Users:
1. A bug was created by the previous update, in calculating optics with
a fraction of a component. This is done in DRAW, for instance, or OFFSET
in MARK. This destructed some parts of memory.
K. Oide
13-FEB-1997 11:07:29.79
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update V1.0.1.2b. Machine errors, parallel processing, etc.
Dear Users:
1. It has been corrected that some elements such as MULT were
ignored by machine-error-commands like DELX, etc. According to this
change, a deck with such commands will not give the same result as
before, because now the same commands gives the different machine errors
even if the SEED is same.
2. The internal logic for parallel processing is updated to use
shared memory instead of file to reduce overhead.
3. Function TrackParticles now run with parallel processing, on acsads.
4. The version number will represent:
V1.0.1.2b
^ ^ ^ ^
| | | |
| | | Small change or bug fix. No change of usage.
| | |
| | Small extension. Minor change in usage.
| |
| A big change in usage, such as 6D optics in FFS.
|
A very big change of the architecture: once in 100 years.
5. Current update plan by myself: This is just a plan.
V1.0.2b: Versatile handling of COUPLE and machine errors.
Clear description of design and actual optics.
V1.1b: 6D optics calculation in FFS.
K. Oide
6-FEB-1997 10:13:24.83
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update V1.0.1b. Individual component as variable, etc.
Dear Users:
1. Now each component with same element name can be used in a matching
independently.
Example: INDEP QF.*
FREE QF.2 QF.3
makes the default variable (initially K1) of QF.* independent from QF (=QF.1),
then use QF.2 and QF.3 as the matching variable. The INDEPENDENT(INDEP)
command is necessary, and only FREE does not work.
The current status of the dependence is shown by ATTRIBUTE(ATTR) or
VARIABLE(VAR) commands.
The dependence of components is reset to the initial by DECOUPLE(DECOUP).
See the SAD web page for more detail.
2. A new command RENUMBER(RENUM) is added to renumber the order of components.
Usage: RENUM component
renumbers the order starting from component. This is useful when the starting
point of a bemline is changed by USE command.
3. SPLIT is now obsolete. Use OFFSET of MARK.
4. Now SAD load module prints out the version number at the beginning. The
current version is 1.0.1b. This may be convenient to identify the version.
K. Oide
20-JAN-1997 14:08:54.21
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. Bug in EMIT.
Dear Users:
1. A bug was created on 1/18 in matrix multiplication in EMIT.
K. Oide
17-JAN-1997 10:36:47.14
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. Reverting the fringe.
Dear Users:
1. I reverted the integration of fringe (half nonlinear - linear - half nonlinea
r) ,
introduced on 11/27/1996, to the original one, (nonlinear - linear at entrance).
The reason is it is nothing but eating CPU, and does not improve the accuracy at
all.
K. Oide
13-JAN-1997 17:29:59.92
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. Bug in transferring EMITX, etc., from EMIT to TRACK.
Dear Users:
1. A bug was found in transferring EMITX, EMITY, EMITZ from EMIT
to TRACK. They were initialized to the values saved in the marker
at the beginning of line when TRACK was called. This bug was created on
10/26/1996.
2. During fixing the bug above, I introduced another bug this afternoon,
causing Segmentation Fault.
K. Oide
11-JAN-1997 06:25:44.29
From: KEKVAX::OIDE
To: @SADUSERS.
CC: OIDE
Subj: SAD Update. ^^^ and $$$, etc.
Dear Users:
1. Now a special notation ^^^ can be used to specify the first component of a
beam line. Thus ^^^+(n-1) means the entrance of the n-th component. Also the
exit of the last component is represented by $$$ instead of ***, though latter
can be used in input.
2. A symbol can be used to specify the keywords of functions Element, LINE,
etc., unless values are not assigned to them. LINE[K1,QF] is same as
LINE["K1","QF"], if no values are assigned to K1 and QF.
3. A bug was found in tracking in CAVI, when both HARM and FREQ are zero.
This caused floating exception.
K. Oide
back to SAD Home Page