What can we know of computational inforgs?

*F. Gobbo & M. BeniniUniv. of L’Aquila & Leeds

CC© BY:© $\© C©

WPI5, Univ. of Hertfordshire,March 27-28, 2013

1 of 17

Introduction

What is computational information?

‘Information’ may be understood in many ways, e.g. assignals, natural patterns or nomic regularities, as instructions, ascontent, as news, as synonymous with data, as power, or aseconomic resource, and so forth [Floridi (2011), 226, ouremphasis].

We will consider only a special class: computational information,i.e., where information is part of a modern Von Neumann’s machine(VNM), being by far the most used computing machine model ever.

3 of 17

Why computational inforgs matter

The digital revolution, which began with the popularization ofthe personal computer around 1980, inaugurated an era in whichpeople appear in consort on an extended network intermixedwith other, non-human information processors, all “inforgs”to use Floridi’s term [Beavers (2011), 264, our emphasis].

In our analysis, we address strictly inforgs with computationalinformation only, i.e., computational inforg (c-inforgs).

4 of 17

How computational inforgs are made

Every c-inforg requires to be programmed by an agent (Ag) bydefinition.

� Ags should be real computer programmers.

� In other words, conscious-less artificial Ags (robots) or biologicalAgs (zombies) are not possible.

for details, see [Gobbo and Benini (2013b)].

5 of 17

The dual nature of computational information

As underlined by the philosophers of Computer Science, programsappear to have a dual nature ([Turner (2013)]):

1. source code, the textual nature, which is human-readable;

2. object code, the binary nature, which is machine-readable;

We apply the system-level-model-structure (SLMS) scheme within themethod of levels of abstraction (LoAs) in order to analyse c-inforgs.

6 of 17

The SLMS scheme and the ontological commitmentsBy accepting an LoA a theory commits itself to the existence ofcertain types of objects. . . constituting the LoA. . . while byendorsing the ensuing models the theory commits itself to thecorresponding tokens [. . . ]

The assumption is that there is no direct knowledge of theintrinsic nature of the entities. . . Once a theory has ontologicallycommitted itself to the structural properties of the system,one is entitled to infer indirectly that, whatever the system andits components (i.e. the objects or relata) may in themselves,they must be such as to allow the theory to model at leastsome of their structural/relational properties.

[Floridi (2011), 349, 351, our emphasis]

7 of 17

Knowledge within c-inforgs

Open vs. closed c-inforgs

A crucial distinction emerge from a property of the VNM-counterpart:

� either source code of the c-inforg is open, i.e., it can be inspectedand modified by the Ag;

� or conversely it is closed – and hence the Ag can be a (power)end-user but not a programmer.

Now, we are going to apply the SLMS scheme in both cases.

9 of 17

Remind of the SLMS scheme

System LoA

Structure Model

attributed to

analysed at

identifies

generates

Figure : The SLMS scheme, adapted from [Floridi (2011), 349]

10 of 17

Open c-inforg: a possible LoE of a programmer

informationtechnology

abstractmachine

source codedesign

patterns

attributed to

analysed at

identifies

generates

Figure : The SLMS scheme applied to a programmer at work

11 of 17

Closed c-inforg: an example of LoE for an end-user

black boxdesktop

application

WYSIWYGmenus &windows

users’ bestpractices

attributed to

analysed at

identifies

generates

Figure : The SLMS scheme applied to a ‘classic’ desktop application use

12 of 17

Reflections on open vs. closed c-inforgs

The possible analyses of open c-inforgs can be more fine-grained thanthe one on closed c-inforgs. In particular:

� from Software Engineering, direct metrics (i.e., they do notdepend on any other attribute, see IEEE Standard 1016) can beapplied to open c-inforgs straightforwardly, being purelyquantitative;

� the analyses on closed c-inforgs cannot be purely quantitative, andtherefore they use indirect metrics, which are morequalitative-oriented.

See the presentation by Marco Benini (here!) for details.

13 of 17

Future directions of work

We aim to apply the SLMS scheme of the method of LoAs to comparequantitative and qualitative properties of inforgs. Sketched ideas:

1. Let’s take the inforg user-of-Facebook as a closed c-inforg;

2. Alice (Ag) prefers to access her Facebook account with hersmartphone; then:

� LoE: she loves to share photos more than blogging;� LoA: there is a different graphic end-user’s interface;� LoO: the combination mobile operating system + app allows her to

perform a definite set of actions and not others;

The description of the Ag Bob, who is keen to access his Facebookaccount with his laptop, is an exercise left to the reader.

14 of 17

Future directions of work

We aim to apply the SLMS scheme of the method of LoAs to comparequantitative and qualitative properties of inforgs. Sketched ideas:

1. Let’s take the inforg user-of-Facebook as a closed c-inforg;

2. Alice (Ag) prefers to access her Facebook account with hersmartphone; then:

� LoE: she loves to share photos more than blogging;� LoA: there is a different graphic end-user’s interface;� LoO: the combination mobile operating system + app allows her to

perform a definite set of actions and not others;

The description of the Ag Bob, who is keen to access his Facebookaccount with his laptop, is an exercise left to the reader.

14 of 17

Future directions of work

We aim to apply the SLMS scheme of the method of LoAs to comparequantitative and qualitative properties of inforgs. Sketched ideas:

1. Let’s take the inforg user-of-Facebook as a closed c-inforg;

2. Alice (Ag) prefers to access her Facebook account with hersmartphone; then:� LoE: she loves to share photos more than blogging;

� LoA: there is a different graphic end-user’s interface;� LoO: the combination mobile operating system + app allows her to

perform a definite set of actions and not others;

The description of the Ag Bob, who is keen to access his Facebookaccount with his laptop, is an exercise left to the reader.

14 of 17

Future directions of work

We aim to apply the SLMS scheme of the method of LoAs to comparequantitative and qualitative properties of inforgs. Sketched ideas:

1. Let’s take the inforg user-of-Facebook as a closed c-inforg;

2. Alice (Ag) prefers to access her Facebook account with hersmartphone; then:� LoE: she loves to share photos more than blogging;� LoA: there is a different graphic end-user’s interface;

� LoO: the combination mobile operating system + app allows her toperform a definite set of actions and not others;

The description of the Ag Bob, who is keen to access his Facebookaccount with his laptop, is an exercise left to the reader.

14 of 17

Future directions of work

We aim to apply the SLMS scheme of the method of LoAs to comparequantitative and qualitative properties of inforgs. Sketched ideas:

1. Let’s take the inforg user-of-Facebook as a closed c-inforg;

2. Alice (Ag) prefers to access her Facebook account with hersmartphone; then:� LoE: she loves to share photos more than blogging;� LoA: there is a different graphic end-user’s interface;� LoO: the combination mobile operating system + app allows her to

perform a definite set of actions and not others;

The description of the Ag Bob, who is keen to access his Facebookaccount with his laptop, is an exercise left to the reader.

14 of 17

An advice for ourselves...

The working philosopher of information cannot

avoid the continuous interplay between

quantitative and qualitative points of views of

information

15 of 17

References

Beavers, Anthony F. (2011), “Historicizing Floridi: The Question of Method, theState of the Profession, and the Timeliness of Floridi’s Philosophy of Information,”in Etica & Politica / Ethics & Politics, XIII, 2, 225–275.

Turner, R. (2013), “Programming Languages as Technical Artifacts,” in Philos.Technol., feb, 1–21.

Floridi, L. (2011), The Philosophy of Information, Oxford: Oxford University Press.

Gobbo, F. and Benini, M. (2013b), “Why zombies can’t write significant sourcecode,” Journal of Experimental & Theoretical Artificial Intelligence, forthcoming.

Gobbo, F. and Benini, M. (2013a), “From Ancient to Modern Computing: A Historyof Information Hiding,” in IEEE Annals of the History of Computing, 4 Feb. IEEEcomputer Society Digital Library. IEEE Computer Society. DOI:10.1109/MAHC.2013.1

16 of 17

Thanks for your attention!

Questions?

For proposals, ideas & comments:

federico.gobbo@univaq.it & marco.benini@uninsubria.it

Download & share these slides here:

http://slidesha.re/14rFzQE

CC© BY:© $\© C© Federico Gobbo & Marco Benini 2013

17 of 17