15 Facts About Programming You Probably Did Not Know | Techila
1268
single,single-post,postid-1268,single-format-standard,ajax_updown_fade,page_not_loaded,,qode_grid_1300,footer_responsive_adv,hide_top_bar_on_mobile_header,qode-content-sidebar-responsive,qode-theme-ver-9.4.2,bridge,wpb-js-composer js-comp-ver-4.12,vc_responsive
 

15 Facts About Programming You Probably Did Not Know

programming-languages-500x500

15 Facts About Programming You Probably Did Not Know

0_Capa.jpg (635×335)

The task of programming is becoming increasingly common, but there are still many facts that people do not know about programmers and programming itself. This post features 15 little known facts about programming.

Like other intellectual activities, the task of programming and how people learn to program computers is well studied. In fact, with more and more people learning to program regardless of language, tool or platform used, it is natural that few people actually know about certain important facts about programming and software development.

From an academic standpoint, the areas of software engineering and education bring forward several very interesting studies obtained by experiments whose results are presented in masters’ and PhD theses. And based on these results I chose the facts mentioned in this post along with appropriate references. By the way, if you got interested to know more about each of the points discussed here I strongly recommend seeking the complete reference for more information.

Based on this context, I will present 15 important facts that, unfortunately, are little known by whom programs. But before, a warning: these facts present results of experimental and empirical research that have specific contexts. What I mean is that there is some room for discussion of the applicability and generalization, but knowing what has already been discovered and studied is important and, at least, can instigate discussion and how close that information is to the reality of the reader.

1) Developers slow to ask for help when facing problems

1_PedirAjudaThis is related to the way people learn how to program; basically, the act of teaching follows the line of learning Mathematics: a little theory, one or two examples and many exercises. This format takes learners to try hard on exercises and, quite often, to solve everything themselves without asking for help. This attitude is not bad and is even recommended, but you need to know to what extent should stop trying and ask for some form of help.

Reference: Andrew Begel and Beth Simon. Novice software developers, all over again. ICER ’08 Proceedings of the 4thinternational Workshop on Computing Education Research, 2008.

2) Programmers have a tendency to report their problems incompletely

2_Erro
This fact is related to Psychology field research. The results indicate that when a person has a problem he/she does not report complete information about the problem, especially when it is responsible directly or indirectly. This result has been confirmed experimentally with programmers and one of the main reasons is the following: to fully report a problem is seen as a sign of weakness that can lead to some kind of judgment of skill and proficiency by whoever is listening to the story. This situation is more common when it comes to a fundamental error committed by novices.

Reference: Shrauger, J.S. and T.M. Osberg. The Relative Accuracy of Self-Predictions and Judgments by Others in Psychological Assessment. Psychological Bulletin, 1981. 90(2): p. 322-351.

3) Developers seek other forms of help before talking to coworkers

3_StatckOverflow2
The fact of communication with other people do not have priority when a programmer needs help again is related to the sense of judging of what other people do when they know the difficulty. However, sites like StackOverflow has flourished exploring this type of behavior by aggregating help in various aspects of communities for developers.

Reference: LaToza, T.D., Venolia, G., and Deline. R. Maintaining Mental Models: A Study of Developer Work Habits. in Proc. ICSE. 2006: IEEE.

4) Progress in programming can be classified into 4 stages

4_Progresso4
The classification of a programmer progress is important to support multiple metrics involved in software development and also help project managers and other professionals to evaluate how good the project is as a whole.

Moreover, it is also important to know in which phase of the progress the developer is to, among other things, offer some kind of help so that he does not spend too much time stuck in a specific task to the point of delaying any deliveries. An interesting classification identified (automatically) four possible states of progress:

a) Complex Programming
b) Making Progress
c) Slow Progress
d) Stuck

Reference: Jason Carter, Prasun Dewan. Design, Implementation, and Evaluation of an Approach for Determining When Programmers are Having Difficulty. ACM Group 2010.

5) Developers find beatable and unbeatable barriers

5_BArreira2
This may seem obvious, but it is very important to be detected, since a programming barrier can lead to serious term, team morale and confidence problems. One of the main difficulties of detecting barriers and classify them is the fact that this information may be subjective. In other words, asking directly to the programmer if he/she is with some beatable or unbeatable barrier already affects the result, as it can not always be sincere. There are also some implications in terms of ego and moral just by identifying this type of barrier on programming.
Reference: Andrew J. Ko et al. Six Learning Barriers in End-User Programming Systems. 2004 IEEE Symposium on Visual Languages – Human Centric Computing.

6) There are 6 types of barriers related to programming

6_Barreira
Besides the classification of beatable and unbeatable programming barriers there is a very interesting study that feature each of the possible programming barriers. To help understand the barriers the researchers point out common phrases in each of the ratings below:

a) Design: “I do not know what the computer should do”.
b) Selection: “I know what to do, but do not know what to use”.
c) Coordination: “I know what to use, but do not know how to combine what I need”.
d) Usage: “I know what to use, but do not know how to use”.
e) Understanding: “I thought I knew how to use X, but it does not do what I expected”.
f) Information: “I understand what happened, but I can not check”.

Reference: Andrew J. Ko et al. Six Learning Barriers in End-User Programming Systems. 2004 IEEE Symposium on Visual Languages – Human Centric Computing.

7) Programmers spend approximately 30% of the time surfing the source code

7_Navegando
People who program know that most of the time relies on a editing source code tool. However, how time is divided between the editing tasks remains unclear from the scientific point of view. According to an important study, it was found that approximately 30% of a programmer working time is not spent editing the text (by including, editing ou deleting), but surfing between multiple files along the source code. The navigation involves research, observation, information gathering, memoring and other activities. That is, you could say that programming is an activity whose third part is just contemplative.

Reference: Andrew J. Ko et al. An Exploratory Study of How Developers Seek, Relate, and Collect Relevant Information during Software Maintenance Tasks. Journal IEEE Transactions on Software Engineering archive Volume 32 Issue 12, December 2006 pp. 971-987.

8) Remote programmer productivity is lower than the productivity of local programmers

8_Fast
This claim about productivity is controversial, especially when routines such as home office, remote working and global software development projects had become increasingly high. Anyway, there are concrete evidences based on several metrics of software that, in fact, remote programmers do not produce as much as programmers working together in the same place.

But it does make sense to think this way if we analyze the other facts of this list, for example, the preference for the lack of communication with other people. In fact, informal communication is a major factor that influenced the results of this research, because asking that hint in the meeting during a coffee break is very important according to what was found alone.

Reference: Herbsleb, J.D., et al. Distance, dependencies, and delay in a global collaboration. In: Proceedings of the 2000 ACM Computer-Supported Cooperative Work conference.

9) Most programmers are white, young & male

9_computer-programmer
This statement about the diversity of programmers did not came just from an academic research, but rather Adda Birnir, who is the founder of the recruitment and selection Skillcrush website. It was introduced in the video “Is CODE the most important language in the world?”.

Nowadays is very common to detach minorities in programming, especially the low number of women. However, as some data show, this is not the only profile that has low representation in programming and this can have serious implications when it comes to create code to applications that must deal adequately with certain groups of users.

Reference: Adda Birnir statement about encoders diversity in “Is CODE the most important language in the world?”.

10) The main error messages, execution times and runtime compilation errors and the average time to solve them

10_compile-time-error2
Error messages are very specific to each language problems and runtime compilation errors. To highlight some cases mention the master’s thesis of Suzanne Marie Thompson, as she looked at a lot of Java programmers in different scenarios and collected many interesting facts about them. The tables below include a bit of history about errors and the average time to correct them.

Although the study focus on a very specific context (learning the Java language) is possible to make a comparison with other scenarios and situations and prove that much of the most common errors occur in different contexts.

TempoMedioResolverProblemasTabelaErrosDeExecucaoTempoMedioResolverProblemas

Reference: Thompson, Suzanne Marie. An exploratory Study of Novice Programming Experiences and Erros. Tese de mestrado defendida em 2004 na inversidade de Victoria, Canadá.

11) The software maintenance consumes more than 50% of the effort

11_legacy-code
Software maintenance involves the manipulation of legacy code. There is a study about effort that shows as a result that the division is not equal between creation and maintenance.

In the study that mention a value of more than 50% of the effort due to software maintenance there is also a great discussion on software evolution towards its maintenance and the necessary tasks for both. Surely is worth taking a look at this reference before making that decision about starting to develop the solution from scratch or working with an existing code base.

Reference: Kemerer C.F. and Slaughter S. An Empirical Approach to Studying Software Evolution, IEEE Transactions on Software Engineering, 25(4), pp. 493-509, 1999.

12) The software maintenance consumes between 40% and 90% of costs

12_Sem-título
One of the main rules of business people says it is much more expensive to get a new customer than to keep an existing customer. However, according to software engineering researches, the reality is somewhat different when it comes to code: to keep the code running through maintenance tasks can cost up to 90% of all project costs.

These statistics are very general and were obtained in a very particular context of the 487 organizations studied for this research (which is from 1980). Certainly there are many factors to consider, but at least there is a starting point for analyzing courses and discuss this topic when talking about software maintenance.

Reference: Lientz, B.; Swanson, E.B. Software maintenance management: a study of the maintenance of computer application software in 487 data processing organizations. Addison Wesley, 1980.

13) The work of software maintenance is divided into 4 basic tasks

13_Sem-título
Still talking about maintenance of source code, a study that has greatly influenced the software engineering community classified by analyzing the results of questionnaires the main practices of software maintenance. Four practices were identified:

a) Improvement: Involve changes in functionality
b) Adaptative: Changes in the environment are adapted to the requirements
c) Corrective: Activities for error correction
d) Preventive: Improvements to avoid future problems

The classification of maintenance practices is very important to assist measurements, organize and track bugs, grouping functionality in new versions and manage the programmers’ work.

Reference: Lientz, B.; Swanson, E.B. Software maintenance management: a study of the maintenance of computer application software in 487 data processing organisations. Addison Wesley, 1980.

Lientz, B.; Swanson, E.B.; Tompkins, G.E. Characteristics of applications software maintenance, Communications of the ACM, Vol. 21, pp.466-471, 1978.

14) Costs of fixing bugs after implementation are 10 times higher than construction phase and 100 times larger than design phase

14_059.TRONSS.32_905
This fact is a classic of technology field and led to the evolution of traditional software development processes until what we have today. The main point here is the identification of high costs when there is a lack of attention to construction and design the software.

Reference: Barry W. Boehm: Software Process Management: Lessons Learned from History. ICSE 1987: 296-298.

15) Peer code review can discover up to 60% of bugs

15_PairProgramming
Code review made by other people, either in the form of pair programming or not, is really effective. There are many studies on this, but one of the key of them indicates that up to 60% of bugs can be discovered (but not necessarily fixed) when more than one person reviews the source code.

This study is relatively old and can be said that it is one of the key influencers of techniques involving agile process and other ways of developing software whose main focus is on activities, steps, organization and other skills not as technical as programming.

Reference: Barry W. Boehm: Improving Software Productivity. IEEE Computer 20(9): 43-57, 1987.

Source: Blog do Mauro Pichiliani

Author: techila

No Comments

Post A Comment