24
Link……? Presented by Eric VITOULEY, trainee. Supervisor : Bram ADAMS

130823 eric vitouley - supporting source code maintenance using screencasts

Embed Size (px)

DESCRIPTION

Screencasts, software developers, social studies, training, support, maintenance

Citation preview

Page 1: 130823   eric vitouley - supporting source code maintenance using screencasts

Link……?

Presented by Eric VITOULEY, trainee. Supervisor : Bram ADAMS

Page 2: 130823   eric vitouley - supporting source code maintenance using screencasts

Two videos… (from www.youtube.com; keyword : firefox)

What is the fundamental difference ?

Page 3: 130823   eric vitouley - supporting source code maintenance using screencasts

But… What is a screencast ?

A screencast is a digital video recording that captures actions taking place on a computer desktop. Screencasts, which often contain voice-over narration, are useful for demonstrating how to use specific operating systems, software applications or website

features.

Page 4: 130823   eric vitouley - supporting source code maintenance using screencasts

3 categories of screencasts

Page 5: 130823   eric vitouley - supporting source code maintenance using screencasts
Page 6: 130823   eric vitouley - supporting source code maintenance using screencasts

Identifying important features…

Virus Scanning

Link Scanner

Theme ChangingMeta-data

Page 7: 130823   eric vitouley - supporting source code maintenance using screencasts

Report Bugs ! Which part Which part of the code of the code

??

Information extraction & Identifiers in source code

Page 8: 130823   eric vitouley - supporting source code maintenance using screencasts

User feedback…

End user software testing

Test/Comparisonscreencasts

Software Comparison

Ideas for next Release

Page 9: 130823   eric vitouley - supporting source code maintenance using screencasts

Keeping documentation up to date…

Documentation

Release v1.0 New Release v1.1

Which part Which part of those doc of those doc

should be should be changed ?changed ?

5s

5s

Page 10: 130823   eric vitouley - supporting source code maintenance using screencasts

So far…

ScreencastsScreencasts

Page 11: 130823   eric vitouley - supporting source code maintenance using screencasts

ScreencastsScreencastsSource Source Code Code

Page 12: 130823   eric vitouley - supporting source code maintenance using screencasts

About traceability…

Extraction

Extractio

n

Page 13: 130823   eric vitouley - supporting source code maintenance using screencasts

Research Questions

Does the audio file yield syntactically correct,

meaningful text ?

Does a screencast file yield syntactically

correct, meaningful text ?

Can we correctly link screencast to source

code ?

Page 14: 130823   eric vitouley - supporting source code maintenance using screencasts

Speech to text…

Dragon Naturally Speaking tool

Filtered by Aspell list

If you handed it Kannel with basic text editors like Notepad you're probably woefully familiar with Google Earth air messages like this parse error on line 8 one set of waiting until you open your file in Google Earth discover errors in your Kannel this video tutorial will show you how to use Janet invalidate your Kannel as you type the first step is to make sure that you have the Java runtime environment installed on your computer you don't have already you can get [email protected] next a J edit.org and install the latest stable version of genetic which is currently 4.2 was the programs on your computer Glenn launch J edit first we need to install some plug-ins to work with XML documents so click on the plug-ins menu plug-ins manager then click uninstall tab click on download options in an update from your list pick a location that's geographically close to you click okay the list of plug-ins will refresh buttons for this alphabetically by clicking on the name column then scroll down to the bottom because were…

IfYouHandedItWithBasicTextEditorsLikeNotepadYouReProbablyWoefullyFamiliarWithGoogleEarth

Syntax correct fo

r majority

of words but n

ot for sentences !

Syntax correct fo

r majority

of words but n

ot for sentences !

Page 15: 130823   eric vitouley - supporting source code maintenance using screencasts

Text from video…

Frames

VoiceAlre

ady done !

Already done !

filtering weird symbols + Aspell list

Name [ Version Status0 Latest Version Check 1.5 Loaded0 OuickNotepad 4.2 LoadedRemoveHelpDownload Options.. Plugin Options... CloseIflPlugin Manager00®Manage Update instaiName Category Version Size□ Calculator Applications 1.1.6 124KB□ Console Applications 4.2.6.5 378KB□ JcrontabPlugin Applications 0.7.2 81KB□ JMailApplications 14 431KB□ JTAPIugin Applications 01 565KB _["I 1 arvfil Rrioftr Annlir.atinns _2_l.fi -7_1 MR.........................................

PlacesMyIIIFfAddSearchMaMMFtftPlacesMyEdAddSHMyOpenTrctr

More words syntactically incorrects

More words syntactically incorrects

greater waste

of

greater waste

of

informatio

n !

informatio

n !

Page 16: 130823   eric vitouley - supporting source code maintenance using screencasts

Images and video processing…

Feature Tracking Contour Detection

Foreground Catching

Page 17: 130823   eric vitouley - supporting source code maintenance using screencasts

Similarity Calculation Process

Text from voice

Text from voice

Text from im

ages

Text from im

ages

FILTERED

FILTERED

WS4J

Calulate similarity with WordNet Similarity tool

Documentation

Array[n x m] words

Page 18: 130823   eric vitouley - supporting source code maintenance using screencasts

Similarity Calculation Process (2)

WORDNET SIMILARITY

WUP

LESK

JCN

PATH

LCH

HSO

LIN

RES

• Threshold : 0,5 Dice Coefficient

Dice Coefficient

Matching average

Matching average

Source : Measuring similarities between sentences, Thanh Ngoc Dao et al.,

Page 19: 130823   eric vitouley - supporting source code maintenance using screencasts

Similarity Calculation Process (3)

Relative similarity values to be able to compare.

Max similaritiy values for each couple of file (X and Y)

A B C

P 1,0 1,0 1,0

N 1,0 1,0 1,0

M 1,0 1,0 1,0 Relative similarity :

Page 20: 130823   eric vitouley - supporting source code maintenance using screencasts

Case Study (1)

Low similarity throughout !

Low similarity throughout !

Page 21: 130823   eric vitouley - supporting source code maintenance using screencasts

Case Study (2)

Low similarity throughout !

Low similarity throughout !

Page 22: 130823   eric vitouley - supporting source code maintenance using screencasts

Work in progress…

Split the screencast into scenes and compute similarities per scene.

Advanced identifier splitting techniques for filtering the video/audio data.

Deeper look on images/video processing to gain more information

Similarity measurement without removing duplicate words

Try to find link directly between screencasts and source code (by splitting identifiers)

Build a model for traceability links retrieval between screencasts and source code

Page 23: 130823   eric vitouley - supporting source code maintenance using screencasts

Conclusion

Page 24: 130823   eric vitouley - supporting source code maintenance using screencasts