CSE BUET Students and Alumni Web
Innovative Projects by CSE BUET students
Bangla SMS, The Journey Continues...
by Sujoy Kumar Chowdhury
Copyright: 3SM Systems, 2005+
The 3SM Systems members at the CSE Micro Computer Laboratory. From left: Shapla, Shihab, Mahbub and Sujoy.
It all started one January 26 night, with an email at the Yahoo Group for BUET ACM contest teams. Elias Mohammad Abu Hasnat, a CSE ’97 alumnus, informed the group that Pacific Bangladesh Telecom Limited (Citycell) was looking for some programmers to introduce Bangla Messaging Services. This email was forwarded to the CSEBUET Yahoo Group by Ragib Hasan (CSE ’96), the owner and moderator of that group.
I called Hasnat Bhai that very night. I discovered that we two had actually worked together as volunteers in the 2003 ACM ICPC venued at BUET. Nice memories. I informed him about my interest in that project, along with my team-mate Nahid Mahfuza Alam (Shapla). Hasnat Bhai told us that Hasan Shihab Uddin and Md. Mahbubur Rahman, another two of my classmates were also interested in that project. He suggested that we form a single group and meet the Citycell AVP for details. So did we.
One fine morning, January 29, we were on our way to the Pacific Centre, the Citycell Head Office at Mohakhali. We met Mr. Ahmed Armaan Siddiqui, the Citycell AVP for Marketing and Value Added Services. He was a nice gentleman and arranged a quick meeting with Marketing Executive Mr. Adnan Bin Taj, Citycell Engineers Mr. Abu Zaid Md. Hussain & Mr. Shahrear Seraj.
We represented ourselves as a team from CSE BUET initially under the banner “xSys Incorporates”. We did some home work prior to that meeting and briefed the Citycell officials about the existing technologies of doing SMS in different native languages. Then we proposed two options for the development of Bangla SMS system: one server- side and another one for the java enabled sets. The officials opted for the server-side solution, primarily for Nokia sets. They had a tight deadline, and wished to launch the system on the next 21st February, the International Mother Language Day. That meant that we had to develop the system prior to 14 February to facilitate enough time for server testing. So hardly two weeks for a corporate solution? Not only that, the officials confirmed us that there had been competing groups, and they would select only the first group with successful output. We took the challenge and responded affirmatively to all of their requirements.
It took us three ‘twenty four hours’ before we produced the first prototype of our system. Yes, three days and three nights, we utilized it all. We sat together and finally settled on a design. Due to the time-constraint as well as competition, we divided the job amongst ourselves. We downloaded heaven and earth right from the emulators to SDKs and specifications, gigabytes of them. This ‘ruthless’ data mining and R&D later proved its point. Co-incidentally, we had strikes at that time, and that saved us the class hours.
We had to fix some rules for the phonetic input method. We chose the rules of Bangla grammar we studied at school, only three in number. Vowel codes are independent, consonants needed both consonant and vowel codes and for Juktakkhors ( conjunct letters), one had to write all the related consonant codes, followed by the appropriate vowel code. The design was such that to an average user the rules were obvious, and the phonetic codes required just two glances to learn them by heart.
The Citycell officials were surprised to see our progress. They initially had a slightly negative notion about the input method. But we convinced them that it was not like a usual PC word processor. Here the rules should completely determine the output, with zero tolerance to ambiguity and error. The three inputs of ‘AMAR’, ‘AAMAAR’ and ‘OMAR’ justified our arguments.
We were yet to design the glyphs for the Juktakkhors when we showed our first prototype. But the next challenge came totally from a different angle. Citycell Engineering informed us that we would have to implement the ESME client also to handshake with the SMSC using the SMPP protocol. Not only that, at that time, Citycell was yet to transfer any binary message officially. So we would also have to study the appropriate PDU format for binary picture messages.
Our initial giga-web-crawling came handy at this point. Without any scope of testing in a real server, we developed the ESME client, along with the picture message PDU, based on our superficial but ever-increasing conception on the two. In this course, we came to know that we had to make official written application to test at the live server. We did. Our development of untested full-featured application was completed by February 11.
But, we were yet to get the clearance from the Citycell Engineering to test at the live server. It was because they had special services featuring the Valentine’s Day (February 14). It was followed by the launch of their another value added service of Group SMS on February 23. And by that time, the pre-scheduled deadline of Bangla SMS on February 21 had already expired! We were very demoralized. Because we had no written confirmation from Citycell that they are going to take our work, let alone an official agreement.
We kept on pinging Citycell, and their verbal assurance kept us alive. Then, one fine morning near March, we got the permission to test live at their server. But alas! Our initial messages produced nothing but garbage for all except for the word ‘Ami’. Readily we sensed that it might have something to do with the dimension of the picture message, which proved right later. We dumped all the binary garbage into a log file and returned home. That very day, while debugging the binary hex data, we figured out that the picture format according to the specification was different than we thought it was. God bless the specification writers! We composed three messages in pure binary according to the assumed specification and sent to the Citycell Engineers via email. They input them directly into their server and compatible sets.
Voila! Success it was, and happy were we. Our system worked! We visited Citycell next morning, viewed the success ourselves. Then we asked the management for an early agreement, because we were dying out of uncertainty and excitement. They assured us.
The logo of 3SM Systems.
Getting their verbal green signal, we bought our domain for banglasms.com on March 4. We changed our group name to 3SM Systems, taking the initials of four of us. Sanjoy designed the beautiful logo for us, with our theme ‘Coders & Creators’. Everything went okay, except for the agreement. Management had a very busy schedule. Unfortunately Mr. Armaan became seriously ill on the day of appointment, and was admitted to hospital for several days. It was followed by our team-mate Shihab’s catching Chicken-Pox. We had to attend our classes, and still we were hopeful about the next suitable date of launch, April 14-The Bengali New Year’s Day.
In the mean time, we delivered our full-featured application to Citycell for testing for seven days, and we were to have our agreement signed on the eighth day, the April 11. Then a great misunderstanding took place. On the eighth day, our application stopped working according to the time limitation, and this postponed our agreement for indefinite period. Finally we convinced Citycell that the expiration of the evaluation period was just part of the business process of Software Engineering and was not at all negatively intentional. On April 26, amidst very intense academic pressure, we were invited to Pacific Centre in a very cordial atmosphere. We signed the agreement, were given an official letter of appreciation and of course, the full payment along with a complimentary Nokia 2280 set with ‘Alaap Call Me’ connectivity. To us, recognition was greater than the remuneration. Citycell decided to re-implement the ESME client on their own following ours due to security issues and required us to submit the deliverable as a library so that they can use it in their future Bangla SMS related services.
The 3 Steps for doing Citycell Bangla SMS.
Citycell launched the Bangla SMS service on May 13 Friday, during our term final exam for Level 3 Term 2. Adrenaline and felicitations are all we remember. According to the agreement, the advertisement in Prothom Alo also contained our logo, group name and link to www.banglasms.com. Although unfortunately, the phonetic rules had not and never been included in the advertisements for Bangla SMS, raising the confusion level amongst potential users.
The days were passing very quickly. The Prothom Alo Onno Alo reporter Ms. Asha Nazneen published an exclusive article on us on 20th May. That was a really big coverage. It was followed by more coverage in more than six leading dailies and airtime in three TV channels. On August 4-6, we solely represented BUET at the Sixth SoftFair 2005 arranged by North South University Computer Club (NSUCC) with our project Bangla SMS at the Bashundhara City Exhibition Centre. We clinched the title of Best University Project. 3SM Systems got selected for the CSE BUET Award 2005 for their outstanding performance and contribution to industry while being in academia.
By this time, there has occurred a huge revolution in the mobile arena of Bangladesh. Amongst other series of events, GrameenPhone and Aktel introduced GPRS, and the tax over mobile sets were reduced in the budget. That resulted in a dramatic change in the availability of feature-rich Java enabled mobile sets in Bangladesh. We were planning to launch our previously developed Bangla SMS solution for mobile sets. This would have enabled the users to send Bangla SMS as text message, not as picture message. But the problem was that both the sender and the receiver would have to have this software installed at their sets. Due to this limitation, we were waiting for an appropriate time to launch. This software was independent of operators. Still we thought without co- operation of the operators and the set manufacturers, we won’t be able to serve the greater user base.
Waiting too much had its payments. Developers from all around the country suddenly got interested in re-implementing Bangla text SMS. I remember Ragib Bhai suggesting one such party in a Yahoo Group to engage in something newer and more innovative. Competition was supposed to be healthy. Until with some very incomplete remote implementations and without any public release, some of them got press and claimed to be the pioneers of Bangla SMS in different occasions, including wrong comments regarding our solutions. That triggered us launch the first free public release of our text based Bangla SMS software for mobile sets.
3SM Bangla Text SMS screenshot.
From August 23, 2005 to this date, this software remains freely available for download at our website www.banglasms.com. This software enables a user to input Bangla message at the rate of merely two keystrokes per letter, which is pretty revolutionary in the context of mobile sets. Moreover, this software is equipped with several intuitive visual input methods; complete Juktakkhor (conjunct letter) support and full featured Bangla SMS inbox & outbox. This software is very lightweight, can be downloaded via GPRS or data cable or even OTA. The software we developed is a generic MIDP 1.0 application, using only one feature of MIDP 2.0 for auto-invoking. Some very interesting features have also been implemented by 3SM Systems which will be available from the next updated versions.
3SM Systems seeks sponsorships from the mobile operators and mobile device manufacturers for the research and development of their free text based Bangla SMS software. This would enable us to render more in future.
We have very big plans regarding Bangla in mobile platform. Few of them remains as plans, and most of them are implemented. It is not limited to Bangla SMS, but extends to complete Bangla interface and many more. According to our professional ethics, we do not like to boast about any work before they are released publicly. And we know introduction of every new technology requires appropriate timing which should be synchronized with the adaptability of the general users. That’s why we do not recommend introduction of manufacturer dependent implementation any more to avoid public confusion, although we have them ready up our sleeves. Because such confusions and mixed standards would delay the standard localization of Bangla in mobile platform, as has happened in the PC platform. Hope one day the general users would evaluate this initiatives properly, and till then our journey continues.
