coyote’s guide to idl programming

Coyote’sGuide toIDLProgramming

David Fanning

Five Ways To Improve Your IDL Programming Now

• Learn the Essential Web Resources• Learn How Colors Work in IDL• Learn 10 Essential Graphics Keywords• Learn the “IDL Way” to Avoid FOR Loops• Learn to Produce Perfect PostScript Output

Six Essential IDL Resources

• Coyote’s Guide to IDL Programming

• IDL Newsgroupcomp.lang.idl-pvwave

http://www.dfanning.com

• Curve Fitting and Mathematicshttp://cow.physics.wisc.edu/~craigm/idl/idl.html

• IDL Emacs Mode (IDLWAVE)http://www.idlwave.org/

• NASA Astronomy Libraryhttp://idlastro.gsfc.nasa.gov/homepage.html

• JHUAPL IDL Libraryhttp://fermi.jhuapl.edu/s1r/idl/s1rlib/local_idl.html

NASA Astronomy Library (http://idlastro.gsfc.nasa.gov/homepage.html)

• Only way to read FITS data• Hundreds of useful, and well-tested routines for

math, statistics, and astronomical utilities• Should already be on your path• The Astronomy Links page is the source of IDL

software used for astronomy

JHUAPL IDL Library (http://fermi.jhuapl.edu/s1r/idl/s1rlib/local_idl.html)

• Solid, well-written library maintained by Ray Sterner• Excellent for time-series data• Date and time routines are best anywhere• Array processing code, filters• Math routines• Often used in conjunction with Astronomy Lib

Craig Markwardt’s IDL Library (http://cow.physics.wisc.edu/~craigm/idl/idl.html)

• Statistical routines • Specialized peak and ellipse fitting applications • Non-linear optimization routines• Optimized set operations (intersect, union, XOR)• Mathematics (array operators, quaternions, etc.)• Integration and differentiation routines• Non-linear least-squares curve fitting

MPFIT Curve Fitting

• Non-linear least-squares curve fitting• Replacement for CURVEFIT• Robust MINPACK-1 FORTRAN subroutine• Unlimited number of parameters• Each parameter can be held fixed or limited to

defined max and min values• Easy access to MPFIT through wrappers• An interactive interface to MPFIT is available

(PAN from NIST)

Using MPFit

PlotErr, t, r, rerr, PSym=3

expr = ‘P[0] + Gauss1(X, P[1:3])

start = [950D, 2.5, 1.0, 1000.0)

result = MPFitExpr(expr, t, r, rerr, start)

OPlot, t, result[0] + Gauss1(t, result[1:3])

Print, result

997.61864 2.1550703 1.4488421 3040.2411

Constraining Parameters

expr = ‘P[0] + Gauss1(X, P[1:3])

pinfo = Replicate( { fixed:0, limited:[0,0], limits:[0.D, 0.D] }, 4 )

pinfo[0].fixed = 1

start = [1000.D, 2.5, 1.0, 1000.0]

result = MPFitExpr(expr, t, r, rerr, start, PARINFO=pinfo)

Print, result

1000.0000 2.1549624 1.4427532 3021.8174

Fix the constant parameter at 1000.

Constraining Parameters

expr = ‘P[0] + Gauss1(X, P[1:3])

pinfo = Replicate( { fixed:0, limited:[0,0], limits:[0.D, 0.D] }, 4 )

pinfo[1].limited[0] = 1

pinfo[1].limits[0] = 2.3

start = [1000.D, 2.5, 1.0, 1000.0]

result = MPFitExpr(expr, t, r, rerr, start, PARINFO=pinfo)

Print, result

997.56563 2.3000000 1.4557194 3035.0424

Limit the mean to a minimum of 2.3.

IDLWAVE Emacs Mode(http://www.idlwave.org/)

• Color-coded editor• Outstanding HTML help• Optional auto-corrected typing for enforcing syntax• Code completion and checking• Several types of interactive debugging tools• Catalog system can scan ALL IDL libraries!• Outstanding support via IDL newsgroup

Six Essential IDL Resources

• Coyote’s Guide to IDL Programming

• IDL Newsgroupcomp.lang.idl-pvwave

www.dfanning.com

• Curve Fitting and Mathematicshttp://cow.physics.wisc.edu/~craigm/idl/idl.html

• IDL Emacs Mode (IDLWAVE)http://www.idlwave.org/

• NASA Astronomy Libraryhttp://idlastro.gsfc.nasa.gov/homepage.html

• JHUAPL IDL Libraryhttp://fermi.jhuapl.edu/s1r/idl/s1rlib/local_idl.html

Coyote’s Guide to IDL Programming

(http://www.dfanning.com)

• 300+ articles in the IDL Tips and Tricks section• Tutorials on essential subjects• 75+ IDL example programs• IDL bug reports• Ask an IDL Question • Order an IDL book• Links to other sites• Learn the latest about Coyote’s adventures

Must-Read Articles for Astronomers

• Dimensional Juggling Tutorial• Array Concatenation Tutorial• Histogram: The Breathless Horror and

Disgust• Are FOR Loops the Embodiment of Pure Evil?• Are FOR Loops Really Evil?• Average Astrophysicist Increased Program

Speed by a Factor of 8100! (http://tinyurl.com/5w3y9)

• How to Get Perfect PostScript Output

IDL Newsgroup(comp.lang.idl-pvwave)

• Friendliest newsgroup on the Internet• 20+ IDL experts share their knowledge• RSI engineers lurk on the newsgroup• Searchable database on groups.google.com• Hitting SEND button is best way to learn IDL• Info on IDL Expert Programmers Association

True-Color Color Decomposition

If your graphic output looks like this youdon’t understand color decomposition.With 24-bit graphics, color decompositionis always ON by default.

IDL> Plot, data, Color=255

IDL> Help, /Device Simultaneously displayable colors: 16777216

Number of allowed color values: 16777216

IDL Color Table Entries: 256

NOTE: this is a TrueColor device

Using Decomposed color

Steel Blue (70,130,180)

1 0 1 1 0 1 0 0 1 0 0 0 0 0 1 0 0 1 0 0 0 1 1 0 = 11829830Lredgreenblue

IDL> Device, Decomposed=1

IDL> Print, 70L + (130L * 2L^8) + (180L * 2L^16)

11829830

IDL> Plot, data, Color=11829830L, Background=‘ffffff’xL

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 = 255Lredgreenblue

IDL> Plot, data, Color=255

01234.

01234..

180255

01234...

Index R G B

132522702...

674487

13038...

80178223180177

01234...

Index R G B

24-Bit Color 8-Bit Color

Plot, data, Color= 11829830L Plot, data, Color=3

Color Decomposition ON Color Decomposition OFF

Steel Blue (70,130,180)

24-bit Non-Decomposed Color

IDL> TVLCT, 70, 130, 180, 240

IDL> Plot, data, Color=240, Background=255

Device Independent Color

• Use FSC_Color for specifying colors by name

IDL> Plot, data, Color=FSC_Color(‘Steel Blue’)

• Color decomposition independent• Works in PostScript, Z-graphics buffer, display• Stay away from top and bottom of color table for drawing colors• 88 colors are available. Can easily add your own colors or load them from a file

Example Code

bg = FSC_Color(‘ivory’)

fg = FSC_Color(‘navy’)

Plot, data, Color=fg, Background=bg, /NoData

Oplot, data1, Color=FSC_Color(‘saddle brown’)

Oplot, data2, Color= FSC_Color(‘indian red’)

Oplot, data3, Color= FSC_Color(‘forest green’)

IDL> Loadct, 13

IDL> TV, image

IDL> Loadct, 13

IDL> TV, image

IDL> TVImage, image

IDL> image = TVRead()

Position Plots with POSITION

Plot, data, Position=[0.10, 0.10, 0.45, 0.90]

Plot, data, Position=[0.55, 0.10, 0.95, 0.65], /NoErase

XYOutS, 0.75, 0.8, ‘Position Plots Here’, Alignment=0.5, $

/Normal, Font=0

Always use Normal or Data coordinatesfor device-independent placement oftext and graphics.

IDL> Plot, data

IDL> Plot, data, Position=Aspect(1.0)

Find a Positionable TV Command

IDL> Pos = [0.1, 0.1, 0.9, 0.9]

IDL> TVImage, image, Position=pos, /Keep_Aspect, /NoInterpolation

IDL> Plot, findgen(100), /NoData, Position=pos, /NoErase

•TVImage•ImgDisp•PlotImage

Avoid Auto-scaling Axes

Plot, data, XRange=[23.4, 78]

Plot, data, XRange=[23.4, 78], $XStyle=1

Avoid Auto-scaling Axes

Contour, image, NLevels=12

Contour, image, NLevels=12 $XStyle=1, YStyle=1

Plotting Symbols

Plot, data, PSym=2

Plot, data, PSym=-2

1 Plus sign (+)

2 Asterisk (*)

3 Period (.)

4 Diamond

5 Triangle

6 Square

8 User-defined.

Create Plotting Symbols(USERSYM)

Filled Circlephi = Findgen(32) * (!PI * 2 / 32.)

phi = [ phi, phi[0] ]

UserSym, Cos(phi), Sin(phi), /Fill

Plot, data, PSym=8, Symsize=1.25

Filled TrianglesUserSym, [ -1, 1, -1, -1 ], [1, 0, -1, 1 ], /Fill ; Filled right-facing triangle.

UserSym, [ 1, -1, 1, 1 ], [1, 0, -1, 1 ], /Fill ; Filled left-facing triangle.

UserSym, [ -1, 1, -1, -1 ], [1, 0, -1, 1 ] ; Open right-facing triangle.

Color with Symbols

phi = Findgen(32) * (!PI * 2 / 32.)

phi = [ phi, phi[0] ]

UserSym, Cos(phi), Sin(phi), /Fill

Plot, data, /NoData

OPlot, data, Color=FSC_Color('firebrick'),

Oplot, data, color=FSC_Color('forest green'), $

PSym=8, Symsize=1.5

Keyword Inheritance (_Extra)

PRO MyPlot, data, Color=color, DataColor=dataColor

IF N_Elements(color) NE 0 THEN $ Message, ‘Color is replaced by DataColor keyword’

IF N_Elements(dataColor) EQ 0 THEN dataColor=‘steel blue’

Plot, data, Color=FSC_Color(dataColor)

PRO MyPlot, data, Color=color, DataColor=dataColor, _Extra=extra

IF N_Elements(color) NE 0 THEN $ Message, ‘Color is replaced by DataColor keyword’

IF N_Elements(dataColor) EQ 0 THEN dataColor=‘steel blue’

Plot, data, Color=FSC_Color(dataColor), _Extra=extra

Keyword Inheritance (_Extra)

IDL> MyPlot, findgen(11), Charsize=2.0, Title=‘My Color Plot’, DataColor=‘green’

extra = { charsize: 2.0, title=‘My Color Plot’ }

IDL> kw = PSConfig(/European)

IDL> Set_Plot, ‘PS’

IDL> Device, _Extra=kw

Avoid FOR Loops if PossibleIDL> array = Indgen(3,4)

IDL> Print, array 0 1 2

9 10 11

Fortran Way:

For j=0,2 Do Begin

For k=0,3 Do Begin

array(j,k) = array(j,k) * 3

Endfor

Multiply array by 3IDL Way:

array = array * 3

Array OperatorsSet all values greater than 5 to 5.

Fortran Way:

For j=0,2 Do Begin

For k=0,3 Do Begin

IF array(j,k) GT 5 THEN array(j,k) = 5

Endfor

IDL Way:

array = array < 5

IDL> array = Indgen(3,4)

IDL> Print, array < 5 0 1 2

WHERE much faster than IFSet all values between 5 to 8 equal = 15.

Fortran Way:

For j=0,2 Do Begin

For k=0,3 Do Begin

IF (array(j,k) GE 5) AND (array(j,k) LT 8) THEN array(j,k) = 15

EndFor

IDL Way:

index = Where((array GE 5) AND (array LE 8), count)

IF count GT 0 THEN array[index] = 15

Convert 1D Indices to 2D Indices

array = Round(Randomu(-3L, 5, 5) * 10)

Print, array 9 6 8 6 1

10 0 2 1 10

9 9 8 2 2

8 1 7 0 1

7 0 2 1 3

indices = Where((array GE 3) AND (array LE 7), count)

IF count GT 0 THEN array[index] = 99

10 0 2 1 10

9 9 8 2 2

8 1 99 0 1

99 0 2 1 3

Convert 1D Indices to 2D Indicesarray = Round(Randomu(-3L, 5, 5) * 10)

10 0 2 1 10

9 9 8 2 2

8 1 7 0 1

7 0 2 1 3

indices = Where((array GE 3) AND (array LE 7), count)

result = Array_Indices(array, indices)

col = Reform(result [0,*])

row = Reform(result [1,*])

IF count GT 0 THEN array[col, row] = 99

10 0 2 1 10

9 9 8 2 2

8 1 99 0 1

99 0 2 1 3

Convert 1D Indices to 2D Indicesindices = Where((image GE 0.55) AND (image LE 0.65), count)

IF count GT 0 THEN BEGIN

result = Array_Indices(image, indices)

col = Reform(result [0,*])

row = Reform(result [1,*])

TV, image

PlotS, col, row, /Device, Color=FSC_Color(‘yellow’)

Dimensional JugglingMultiply each column of array by a vector, b.

Fortran Way:

var = IntArr(3,4)

For j=0,2 Do Begin

var[j,*] = array[j,*] * b

EndFor

IDL> array = Indgen(3,4)

IDL> b = Fix(RandomU(-1L, 4) * 12

IDL> Print, array

0 1 4 4 5 6

10 11 12

IDL> Print, b 4 1 9 6

Dimensional JugglingMultiply each column of array by a vector, b.

IDL> Print, Reform(b, 1, 4) 4

IDL> Print, Rebin( Reform(b, 1, 4), 3, 4) 4 4 4

IDL Way:Print, var = array * Rebin( Reform(b, 1, 4), 3, 4 )

col = Transpose(b)

col = Rotate(b,1)

col = 1 # b

col = b ## 1

Dimensional Juggling Can extend this to any number of dimensions.

IDL> array = Indgen(3,4,3) 0 1 2

9 10 11

12 13 14

15 16 17

18 19 20

21 22 23

24 25 26

27 28 29

30 31 32

33 34 35

IDL> Print, Rebin( Reform(b,1,4,1), 3, 4, 3) 4 4 4

Array ConcatenationIDL> a = Make_Array(4, 4, Value=1B)

IDL> Print, a 1 1 1 1

1 1 1 1

IDL> b = Make_Array(4, 4, Value=2B)

2 2 2 2

IDL> Print, [a, b] 1 1 1 1 2 2 2 2

1 1 1 1 2 2 2 2

Array ConcatenationIDL> Print, [ [a], [b] ]

1 1 1 1

2 2 2 2

IDL> Help, [ [a], [b] ] INT = Array[4, 8]

IDL> Print, [ [ [a] ], [ [b] ] ]

1 1 1 1

2 2 2 2

IDL> Help, [ [ [a] ], [ [b] ] ] INT = Array[4, 4, 2]

Array Concatenation

IDL> b = (Indgen(4) + 1) * 4

IDL> Print, [ [a], [b], [b] ] 1 1 1 1

1 1 1 1

4 8 12 16

IDL> Help, [ [a], [b], [b] ] INT = Array[4, 6]

Add rows to an array.

IDL> b = Reform((Indgen(4) + 1) * 4, 1, 4)

IDL> Print, [ b, a, b ] 4 1 1 1 1 4

8 1 1 1 1 8

12 1 1 1 1 12

16 1 1 1 1 16

IDL> Help, [ b, a, b ] INT = Array[6, 4]

Add columns to an array.

Array Concatenation

IDL> image24 = [ [ [image_1] ], [ [image_2] ], [ [image_3] ] ]

IDL> Help, image24 INT = Array[400, 600, 3]

Create a true-color image

Convert band interleaved image to pixel interleaved

IDL> image24 = Transpose(image24,[2,0,1])

IDL> Help, image24 INT = Array[3, 400, 600]

Index Manipulation

img1 = BytScl(Loaddata(4), Top=99)

img2 = BytScl(Loaddata(5), Top=99)+100B

LoadCT, 13, NColors=100

LoadCT, 3, NColors=100, Bottom=100

index = Where((Indgen(256L*256L) MOD 2) EQ 0)

img1[index] = img2[index]

Window, XSize=256, YSize=256

TV, img1

Quick look at two images simultaneously.

Index Manipulation

IDL> data = Indgen(8,4) + 1

IDL> Print, data 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

17 18 19 20 21 22 23 24

25 26 227 28 29 30 31 32

IDL> evenRowIndex = (Indgen(4/2) * 2) + 1

IDL> data[*, evenRowIndex] = Reverse(data[*, evenRowIndex])

IDL> Print, data 1 2 3 4 5 6 7 8

16 15 14 13 12 11 10 9

17 18 19 20 21 22 23 24

32 31 30 29 28 27 26 25

Reverse the even rows in a 2D array.

Histogram Magic

IDL> a = Indgen(10)

IDL> Print, Histogram(a, Binsize=5) 5 5

IDL> Print, Histogram(a) 1 1 1 1 1 1 1 1 1 1

IDL> data = Randomu(-3L, 100)

IDL> Print, Histogram(data, Binsize=0.20) 25 17 22 18 18

A histogram is used to count how many elements fall into “bins”.

Histogram MagicIf binning data was all Histogram did, it would beuseful, but not spectacular! Think of Histogram as an index manipulator and it becomes a lot more exciting.

IDL> data=fix(randomu(101,25)*12) & Print, data 5 2 6 11 7 1 4 0 10 11 4 0 1 10 0 4 3 11 9 4 1 1 11 4 3IDL> h = Histogram(data, REVERSE_INDICES=ri)

3 4 1 2 5 1 1 1 0 1 2 4 | | | | |6 | | | | | | | || |5 | | |10 | | | | | | |3 ||7 |12 | | |15 | | | | | | |9 ||11 |20 | |16 |19 | | | | | |8 |17 ||14 |21 |1 |24 |23 |0 |2 |4 | |18 |3 |22 |+---+---+---+---+---+---+---+---+---+---+---+---+ 0 1 2 3 4 5 6 7 8 9 10 11

Histogram MagicHistogram is fast, much faster than Where. 3 4 1 2 5 1 1 1 0 1 2 4 | | | | |6 | | | | | | | || |5 | | |10 | | | | | | |3 ||7 |12 | | |15 | | | | | | |9 ||11 |20 | |16 |19 | | | | | |8 |17 ||14 |21 |1 |24 |23 |0 |2 |4 | |18 |3 |22 |+---+---+---+---+---+---+---+---+---+---+---+---+ 0 1 2 3 4 5 6 7 8 9 10 11

The “I” vector The “O” vectorri = iiiiiiiiiiiiiiiiiiioooooooooooooooooooooooooooooo |-----------------||----------------------------| |0 nh||nh+1 nh+total(h)|

The o-vector contains the data indices, in order, and the i-vector just shows us where to go to get them.

| | | | |6 | | | | | | | || |5 | | |10 | | | | | | |3 ||7 |12 | | |15 | | | | | | |9 ||11 |20 | |16 |19 | | | | | |8 |17 ||14 |21 |1 |24 |23 |0 |2 |4 | |18 |3 |22 |+---+---+---+---+---+---+---+---+---+---+---+---+ 0 1 2 3 4 5 6 7 8 9 10 11

IDL> Print, ri[ri[4]:ri[5]-1] 6 10 15 19 23

IDL> Print, data[ri[ri[4]:ri[5]-1]] 4 4 4 4 4

IDL> if ri[4] ne ri[5] then Print, data[ri[ri[4]:ri[5]-1]] else Print, 'No data in bin 4‘ 4 4 4 4 4

IDL> if ri[8] ne ri[9] then Print, data[ri[ri[8]:ri[8]-1]] else Print, 'No data in bin 8' No data in bin 8

Reverse_Indices looks ugly, but works great!

The “I” vector The “O” vectorri = iiiiiiiiiiiiiiiiiiioooooooooooooooooooooooooooooo |-----------------||----------------------------| |0 nh||nh+1 nh+total(h)|

Using Histogram

Find the intersection (common elements) of two vectors.

IDL> sd = -3LIDL> a = Fix( Randomu(sd, 8) * 20 )IDL> b = Fix( Randomu(sd, 8) * 20 )IDL> Print, a, b 17 11 15 11 1 19 0 4 2 19 17 18 16 3 4 16IDL> Print, Where(Histogram(a, OMin=om) gt 0 AND Histogram(b, Min=om) gt 0) + om 4 17 19

Find the union (elements in either vector) of two vectors.

IDL> Print, Where(Histogram([a,b], OMin = om)) + om 0 1 2 3 4 11 15 16 17 18 19

Using Histogram

Remove elements, listed in random order, from a vector.

IDL> vec = Randomu(sd,10) IDL> remove = [3,7,2,8] IDL> keep = Where(Histogram(remove,MIN=0,MAX=N_Elements(vec)-1) eq 0,cnt) IDL> if cnt ne 0 then vec = vec[keep] IDL> print,keep 0 1 4 5 6 9

Find the median value for each quartile of a data set.

IDL> data = Randomu(sd,100)*100IDL> h = Histogram(data,Binsize=25, Reverse_Indices=ri)IDL> med = FltArr(4)IDL> for j=0L,3L do if ri[j+1] gt ri[j] then med[j] = Median(data[ri[ri[j]:ri[j+1]-1]])IDL> print,med 15.2426 40.9219 63.8255 86.1637

Set up Window on PostScript Page

Set_Plot, ‘PS’Device, XSize=xs, YSize=ys, XOffset=xoff, YOffset=yoff, /Landscape

Create a PostScript Window with the Correct Aspect Ratio

Plot, data

keywords = PSWindow()

Help, keywords, /Structure XSIZE FLOAT 5.95000 YSIZE FLOAT 5.95000

XOFFSET FLOAT 1.27500

YOFFSET FLOAT 2.52500

INCHES INT 1

PORTRAIT INT 1

LANDSCAPE INT 0

thisDevice = !D.Name

Set_Plot, ‘PS’

Device, _Extra=keywords

Plot, data

Device, /Close_File

Set_Plot, thisDevice

Plot, data, Position=[0.10, 0.10, 0.45, 0.90]

Plot, data, Position=[0.55, 0.10, 0.95, 0.65], /NoErase

XYOutS, 0.75, 0.8, ‘Position Plots Here’, Alignment=0.5, $

/Normal, Font=0

Always use Normal or Data coordinatesfor device-independent placement oftext and graphics.

Use a Positionable TV Command

IDL> Pos = [0.1, 0.1, 0.9, 0.9]

IDL> TVImage, image, Position=pos, /Keep_Aspect, /NoInterpolation

IDL> Plot, findgen(100), /NoData, Position=pos, /NoErase

•TVImage•ImgDisp•PlotImage

PostScript Setup

keywords = PSConfig(Cancel=cancelled)

IF cancelled THEN RETURN

thisDevice = !D.Name

Set_Plot, ‘PS’

Device, _Extra=keywords

Plot, data

Device, /Close

Set_Plot, thisDevice

Load Drawing Colors Properly

Use PostScript or True-Type Fonts

Plot, data, Font=0

!P.Font = 0

Device, Set_Font=‘Arial*14’, /TT_Font

Plot, data, Font=1

XYOutS, 0.75, 0.8, ‘Position Plots Here’, /Normal, Font=1

You will have better looking plots if you use PostScript or True-Type fonts. If youuse TeXtoIDL be sure to set the PostScriptkeyword.

Protect Device-Specific Code

IF (!D.Flags AND 256) NE 0 THEN $ ; Windows supportedWindow, XSize=400, YSize=400

Other commands to protect:• WSet, wid• Device, Decomposed=0

Coyote’sGuide toIDLProgramming

David Fanning

coyote’s guide to idl programming

idl programmingidl

idl way

idl libraries

htmljhuapl idl libraryhttp

idl programming nowlearn

source of idl software

idl emacs mode idlwavehttp

astronomy libraryhttp

Documents

file: idl-simple-manual.txt = introduction to idl basics

how to quick setup idl- 24/4802 to annex m? applied model:...

corba idl 1 introduction to corba idl overview omg idl is...

idl meteo intro

idl primer’s course

update: idl & envi

idl preprocessor guide -...

user interface programming - department of geological and...

object programming - michigan technological university...

comparing idl to c++ with idl to c++11

definition and utilisation of omg idl to ttcn-3 · pdf...

introduction to idl -...

geog 820.03 basic skills in scientific programming syllabus,...

idl idl_2008_3

netreural idl

coyote’s guide to idl programming david fanning. five ways...

idl basics.pdf

idl classic

introduction to idl dr james scuffham. what is idl?...

idl presentation1