soc.culture.pakistan

Long-form reference · 8,363 words · about a 36-minute read

The group's own paperwork

Very little of what soc.culture.pakistan actually said survives anywhere this directory can reach, but a great deal survives of how it came to exist. The Internet Systems Consortium keeps a mirror of news.announce.newgroups, the moderated group in which every Big-Seven proposal had to be published, and for this group the file is complete. It contains a call for discussion, a call for votes, an interim acknowledgement of ballots received, a final tally naming every voter, and — eleven months after the group opened — a proposal to put a moderator over it. Some forty-five kilobytes of plain text, all of it dated, none of it written with an eye on posterity, which is generally what makes such things useful later.

The proposal was posted on 6 February 1990 by Ali Raja, writing from a Texas Instruments machine, and its reasoning was almost aggressively practical. Discussions of Pakistani matters, it observed, were usually carried on in soc.culture.indian; a number of issues relevant to Pakistan were not relevant there; and the sheer number of articles seemed to justify a group of its own. Discussion was to run a fortnight, until at least 21 February. The header of that first article is itself a map of where the argument had been living up to then: it was cross-posted to soc.culture.indian, soc.culture.turkish, soc.culture.arabic and talk.politics.mideast, with follow-ups directed to soc.culture.indian, which the proposer noted drily had already seen more than its fair share of such discussions.

The call for votes followed on 1 March 1990, issued by Junaid Ahmed Zubairi at Syracuse University, who acted as vote-taker. It specified a world-wide, unmoderated group, and it set out the intended subject matter as a list. That list is the nearest thing the group ever had to a charter of its own composition, and it is worth reproducing exactly, because almost every section below is an entry on it. The proposed group was to cover aspects of Pakistani:

• culture, history, philosophy, ideology, geography;
• societies, traditions, customs;
• literature, poetry, art, folklore;
• languages, books;
• science, technology;
• food, cookery;
• local events, news, programs, economy;
• communities abroad, problems, needs; and
• things normally discussed in the soc.culture newsgroups.

The mechanics were those of the period and are described in full on the branch's own hub page rather than repeated here: votes by e-mail only, a fixed window, no proxies, no ballots posted to the network, and a passing rule requiring a hundred more yes votes than no votes and a two-thirds majority of valid ballots. The hierarchy this group was joining had been assembled during the Great Renaming of 1987, when Usenet's national and regional groups were gathered under a common prefix; the machinery that admitted new members to it is set out at soc.culture.

The vote, and the roll it left behind

Voting ran from 26 February to midnight on 18 March 1990. On 12 March the vote-taker posted a mass acknowledgement of ballots received so far — a courtesy of the era, intended to let voters check that their mail had arrived — and on 19 March he posted the result, under a subject line that left nothing to interpretation: the group passed with 365 yes votes to 71 no. Five further ballots were declared invalid, one for arriving late and four for containing no text and no indication in the header of which way they were meant to be counted.

The arithmetic was set out in the posting itself, in the manner of a returning officer who expects to be checked. There were 294 more yes votes than no votes, against a requirement of 100. Two-thirds of the 436 valid ballots is 290, and there were 365 in favour. Both conditions were met and the result was declared. Then, as the guidelines required, the vote-taker printed the entire electorate: 441 numbered lines, each carrying an address and, where the voter had supplied one, a name.

That roll is the single most informative document the group produced, because it is a dated census of where the readership physically sat in March 1990. Counting the addresses on it gives a distribution that is not open to much argument: 327 at hosts in .edu, 66 in .com, fifteen on BITNET, seven in .ca, five in .gov and five in .uk, with a handful under .org, one in Finland, and the remainder written as ARPA or UUCP-style relay paths. The universities named run from Duke, Columbia and MIT to Purdue, Cornell, Michigan State and Ohio State, and across the Atlantic to Edinburgh, Essex and Sussex. Not one address on the roll ended in .pk. The country-code domain for Pakistan did not yet exist.

A ballot roll is not a readership survey, and this one should not be pressed further than it will go: people who voted were, by definition, people sufficiently invested to send mail to a stranger in Syracuse, and quiet readers leave no trace in it at all. But as a lower bound on the shape of the audience it is unambiguous. In 1990 a group about Pakistan was, of necessity, a group of people outside Pakistan — students, graduate researchers and engineers at institutions that happened to carry a news feed, which in 1990 meant North American and Western European universities and a short list of technology companies.

A description line that outlived everything else

The one-line description that the network files carry for a group is a small thing, but it is remarkably durable, and in this case it has outlived the moderators, the sub-groups and the traffic. The ISC newsgroups file — the canonical list of Big-8 groups and their descriptions — still gives the entry as: soc.culture.pakistan, "Topics of discussion about Pakistan." The corresponding active file still marks it "y", meaning postable.

That wording is older than most of the group's history. The newgroup control message that actually created the group in March 1990 predates the surviving control archive, but the earliest copy the archive does hold — a local re-issue sent on 23 May 1991 from a machine at Compass, Inc. of Wakefield, Massachusetts, one of the housekeeping messages news administrators sent to make sure their own servers had the group — carries that exact sentence as its body. It has been the group's official description, so far as the record shows, for at least thirty-five years. It is also a line that would fit almost any of the hundred-and-thirty-odd soc.culture groups still carried in that file.

The gap between that sentence and the nine-line list of subjects in the 1990 call for votes is the gap this article exists to fill.

A proposal to moderate, February 1991

Eleven months after the group opened, on 21 February 1991, a call for discussion was posted by Muhammad Javed Iqbal, writing from the computer centre of Middle East Technical University in Ankara, proposing that soc.culture.pakistan acquire a moderator. It was cross-posted to news.announce.newgroups and to the group itself, and it went further than most such proposals: it offered a draft rulebook of eight clauses.

Those clauses are worth noting because they show what the group's own readers thought its problems were, barely a year in. Articles blasphemous to any religion were to be rejected, and copies forwarded to the postmasters and administrators of the originating site. Follow-ups to any one article were to be capped — twenty-five was the number suggested. A topic was to be closed to further discussion ten days after the original posting, except in exceptional cases. Meaningless articles and follow-ups were to be rejected outright. Nobody was to be allowed to cross what the author called decency limits, and such articles were to be edited or refused. And two disclaimers: that a moderator's acceptance of an article implied no agreement with it, and that reposting someone else's article implied no agreement either.

The proposal invited amendment on every point, including how many moderators there should be and who they ought to be. It never reached a vote. The Big-8 Management Board's own summary of the group's history, written sixteen years later, records the episode in a single sentence: there was a proposal to convert the group to a moderated status in February 1991, but it never proceeded to a vote. soc.culture.pakistan remained unmoderated for its entire life, which meant that anyone whose site carried it could post to it, and that no copy of it anywhere was more official than any other.

The country, as a dated record: 1947 to 1971

A national newsgroup inherits its subject matter, and the subject matter here is unusually well dated. What follows is chronology and nothing else. Every question in it was argued in the group, most of them repeatedly; none of them is adjudicated here, and the register of this page is deliberately that of a calendar rather than a verdict.

The partition of British India was set out in the Indian Independence Act 1947 and took effect at midnight on 14–15 August 1947, dividing the provinces of Bengal and the Punjab on district-wide majorities and dividing with them the army, the navy, the civil service, the railways and the treasury. The Dominion of Pakistan was established on 14 August 1947, the Dominion of India on 15 August. Within weeks the two dominions were at war over the princely state of Jammu and Kashmir, in the conflict of 1947–48 that is also called the first Kashmir war. A second war followed in August and September 1965, closed by a ceasefire on 23 September and by the Tashkent Declaration of 10 January 1966.

The third war is the one the group returned to most often, and it is the one that changed the shape of the country. The Bangladesh Liberation War began on the night of 25 March 1971 with the Pakistan Army's Operation Searchlight in East Pakistan. India entered the conflict directly on 3 December 1971; Pakistan's Eastern Command capitulated at Dhaka on 16 December 1971, and East Pakistan became the People's Republic of Bangladesh. Pakistan did not formally recognise Bangladeshi sovereignty until February 1974. The Simla Agreement between India and Pakistan was signed on 2 July 1972. The scale of the events of 1971, their naming and their responsibility are all contested, and this page documents that they were contested rather than settling any part of them.

Cricket, which will reappear below, arrived earlier than any of this: Pakistan was granted Test status at a meeting of the Imperial Cricket Conference at Lord's on 28 July 1952, on India's recommendation, and played its first Test at Delhi that October.

Constitutions, and the alternation of government

The constitutional record is equally datable and was equally contested. The Dominion ended on 23 March 1956, when the Constituent Assembly adopted the country's first constitution and established an Islamic Republic; that document lasted two and a half years and was abrogated under martial law on 7 October 1958 by President Iskander Mirza, who was himself deposed by the commander-in-chief, Muhammad Ayub Khan, on 27 October. A second constitution, drafted under and promulgated by Ayub Khan, was in force from 8 March 1962; it established a strong presidential system in which the president held all executive authority, and it was abrogated on 25 March 1969 by Yahya Khan. The third and current constitution was drafted by the government of Zulfikar Ali Bhutto, approved on 10 April 1973 and brought into force on 14 August 1973.

Zulfikar Ali Bhutto's government was overthrown on 5 July 1977 by the chief of army staff, Muhammad Zia-ul-Haq, who ruled until his death in an air crash in August 1988 and remains the country's longest-serving de facto ruler. Elections followed later that year, and Benazir Bhutto became prime minister — the first woman elected to head a democratic government in a Muslim-majority country — in December 1988. That is where the group's own timeline begins: soc.culture.pakistan opened fifteen months into her first administration.

What followed, through the decade in which such groups were most used, was a rhythm of dismissals under the Eighth Amendment, which allowed a president to dissolve the National Assembly. President Ghulam Ishaq Khan dismissed Benazir Bhutto's government in August 1990, five months after the group was created. He dismissed Nawaz Sharif's government in April 1993. President Farooq Leghari dismissed Benazir Bhutto's second government on 5 November 1996. Three heads of government were removed by presidential order within six and a half years, each removal followed by an election. The Eighth Amendment itself was repealed in 1997.

The group's own years, 1990 to 2007

The second half of the group's active life supplied a sequence of events that arrived on the network faster than they arrived in most newspapers, which is a large part of what the room was for.

India conducted five nuclear tests at Pokhran on 11 and 13 May 1998, under the collective name Operation Shakti. Pakistan conducted five simultaneous underground tests at the Ras Koh Hills in Chagai District, Balochistan, on 28 May 1998, code-named Chagai-I, and a further test in the Kharan Desert on 30 May, Chagai-II. Seventeen days separated the first Indian detonation from the first Pakistani one, and both governments' announcements were carried into the newsgroups within hours.

The Kargil conflict was fought between India and Pakistan in the Kargil district of Ladakh and along the Line of Control from May to July 1999. On 12 October 1999 the government of Prime Minister Nawaz Sharif was removed in a bloodless coup led by the chairman of the joint chiefs and chief of army staff, General Pervez Musharraf; on 14 October the constitution was suspended by provisional order. Musharraf served as president from 2001 to 2008. The events of September 2001 and their consequences for the region fell squarely inside the group's remaining active years.

The longer-running dispute behind much of this — Kashmir — was the subject that most reliably produced cross-posted exchanges with soc.culture.indian, and that produced groups of its own: soc.culture.indian.jammu-kashmir was voted on in the same season as the Pakistan branch's reorganisation, and the moderator of news.announce.newgroups mentions the two proposals in the same breath in an April 1996 posting. The paperwork for that one group runs to more than a megabyte in the ISC archive, which is a rough but honest index of how much argument a single newsgroup proposal could generate.

Urdu, and the languages of the provinces

Language is the strongest distinct material this group has, and the position is more complicated than the phrase "an Urdu-speaking country" suggests. Urdu is Pakistan's national language under Article 251 of the constitution and shares official status with English; it is the country's lingua franca and the language of inter-communication between ethnic groups. It is not, however, the first language of most Pakistanis. At the 1998 census — the count contemporary with the period in which such groups were most used — Urdu was reported as the mother tongue of 7.57 per cent of the population. Punjabi was reported by 44.15 per cent, Pashto by 15.42, Sindhi by 14.1 and Saraiki by 10.53. Balochi, Hindko and Brahui follow, and Pakistan has over seventy languages spoken as first languages, of which roughly sixty have fewer than a million speakers each.

This mattered to the newsgroup in a specific way. A national group in a country whose national language is a second language for most of its citizens, and whose readership abroad was drawn from every province, is a group with no single obvious medium. Every surviving document produced by or about the group — the 1990 call for votes, the 1995 charters, the 1996 ballot rolls, the 2007 removal notices — is written in English. That is not only a matter of convenience: until well into the group's second decade, writing Urdu on a computer at all was a specialist activity.

The reorganisation proposal of October 1995 tried to give this a room of its own. It listed soc.culture.pakistan.linguistics — "Pakistani languages & literature" — with a charter that named Urdu, Sindhi, Punjabi, Pashto, Balochi and Saraiki, invited discussion of prose, poetry and ghazals, and then added a clause that reads today like a small window onto the state of the art: "Those who wish to work on text editors & dictionaries for Pakistani languages are welcome to coordinate their efforts in this newsgroup." The linguistics group was dropped before the vote; it does not appear in the revised proposal of November 1995 and it was never created. But the sentence survives, and it says plainly that in 1995 an Urdu text editor was still something a newsgroup might reasonably organise volunteers around.

Nastaliq: the hand the machines could not draw

Urdu is written in the Perso-Arabic script, but the style matters more than the script here. Urdu is conventionally set in Nastaliq, a calligraphic hand that developed in Iran out of a combination of Naskh and Ta'liq, and whose emergence is traced by different scholars to the thirteenth and fourteenth centuries. Arabic itself is normally set in Naskh; Urdu, Persian, Kashmiri, Punjabi in Shahmukhi, Sindhi, Saraiki and Balochi are commonly set in Nastaliq. The difference is not cosmetic. Nastaliq runs on a steep diagonal from upper right to lower left within each word, its letters change shape according to what they join, and many letters take more than three general forms, even in undecorated everyday documents. A line of Nastaliq is closer to a sequence of designed ligatures than to a string of glyphs set side by side.

This is why Nastaliq was, for decades, genuinely hard to produce on a computer, and the record of that difficulty is well documented. Modern Nastaliq typography effectively begins with Noori Nastaliq, first produced as a digital font in 1981 through a collaboration between the calligrapher Ahmed Mirza Jamil and Monotype. It worked by brute force: over twenty thousand individually designed ligatures. It produced accurate results, and it let a newspaper such as Pakistan's Daily Jang replace a room of calligraphers with digital typesetting. But through the 1990s it carried two disabling limitations. It was not available on standard platforms — not on Microsoft Windows, not on the Mac — and text entry was not WYSIWYG: a document had to be composed as a series of commands in Monotype's proprietary page description language.

In 1994 InPage Urdu appeared, a page-layout program for Windows built for Pakistan's newspaper industry by the Indian software house Concept Software, which licensed Noori Nastaliq from Monotype. InPage became the standard tool of Urdu publishing, and it remains in use; it did not become Unicode-based until 2009. Native Nastaliq support in a mainstream operating system came later still. Windows 8 was the first version of Microsoft Windows to ship it, through Microsoft's Urdu Typesetting font. Google's open-source Noto Nastaliq Urdu has been carried by Apple on macOS since High Sierra and on iOS since iOS 11, which is to say 2017 — a decade after this group's sub-hierarchy was dismantled.

It is worth being precise about what the obstacle was, because it is easy to state it wrongly. Nastaliq is not separately encoded in Unicode and never has been; it is a style of the Arabic script, not a writing system, and Nastaliq letterforms are produced by choosing a Nastaliq font. The characters were never the problem. The rendering was.

What the encodings did, and when

The encoding record supports that distinction closely, and it belongs here rather than on the Arabic page, which owns the more general story of Arabic script on a seven-bit wire and is not retold at soc.culture.arabic.

The eight-bit standard for the Arabic script, ISO/IEC 8859-6, was first published in 1987. It encodes only nominal letters, with no preshaped forms, so shaping has to be done at display time — and, decisively for Urdu, it does not include the extra letters needed to write most Arabic-script languages other than Arabic itself. The retroflex and aspirate letters that Urdu requires are simply absent from it. Microsoft's Windows-1256 code page did cover Arabic-script languages including Persian and Urdu, encoding abstract letters rather than shaped glyphs; but it was a vendor code page rather than a standard, it was not compatible with ISO/IEC 8859-6, and a document encoded in it was only as portable as the assumption that the reader was running Windows.

Unicode settled the character question early. The Unicode Consortium was incorporated on 3 January 1991 and the first volume of the standard was published that October. The Arabic block, U+0600 to U+06FF, was there in version 1.0.0 with 169 characters, and those 169 already covered the ranges U+0670–U+06B7, U+06BA–U+06BE, U+06C0–U+06CE and U+06D0–U+06D5, which between them carry every additional letter the Urdu alphabet requires beyond the Arabic one. The Arabic Presentation Forms blocks, which encode the contextual forms and ligatures needed for Persian, Urdu, Sindhi and Central Asian languages, were settled into their present shape in the 1.0.1 and 1.1 revisions. So by the time soc.culture.pakistan was a year old, a standard existed in which Urdu could be represented unambiguously. What did not exist, for years afterwards, was software that would draw it, mail that would carry it reliably, or a keyboard most readers could type it on.

Pakistan's own standardisation effort is dated and public. In 1998 the National Language Authority, under Attash Durrani, began a research and development project to standardise Urdu encoding; the result was Urdu Zabta Takhti, or UZT. Version 1.01 was standardised in July 2000 for electronic computing, communications and storage, and, on the Authority's own account of the project, Urdu language support was incorporated into Unicode versions 3.1 and 4.0 on the basis of that version. The Authority finalised version 1 of the corresponding keyboard on 14 December 1999, and in 2001 the National Database and Registration Authority adopted it for data entry on the computerised national identity card. Every one of those dates falls inside the group's active life, and the last of them falls after the sub-hierarchy's moderated rooms had already gone quiet.

The practical consequence: English, and Urdu in Latin letters

The consequence for any newsgroup running on that wire is easy to state. Latin letters crossed it reliably; the Urdu script did not, and could not be assumed to render at the far end even where a sender managed to produce it. What Urdu speakers did about that is well documented in general, and what follows is that general record. Nothing here is an assertion about this group's own traffic, of which nothing survives in this directory.

This was not peculiar to Usenet. Most Internet use in Pakistan has remained in English; Urdu-language newspapers have maintained Urdu web presences, but common usage online has long been in romanised Urdu, a practice documented in the Pakistani press from the early 2000s onwards. An Urdu scholar writing in 2004 described Roman Urdu as strongly opposed by lovers of the traditional script and used nonetheless by most people on the Internet, precisely because most technologies did not have the Urdu script; he noted that the Jang Group had devised its own romanisation scheme, and that chat services were functioning as laboratories for an evolving orthography.

Roman Urdu has no standard spelling. Users imitate English orthography, which is not itself reliably phonetic; the same word appears in half a dozen spellings, sometimes within one thread. It differs from the informal romanised Arabic used elsewhere on the network in one visible respect — it does not press digits into service as letters, and rarely uses mixed case, because the Arabic letters without a clear Latin equivalent are in Urdu either silent or pronounced identically to letters that do have one. The neighbouring case of a South Asian language transliterated for a Latin-only wire, and what that does to an archive, is treated at soc.culture.bengali and is not repeated here.

The effect on a literary tradition carried this way is double-edged. A romanised couplet is legible on any terminal anywhere, but because Roman Urdu has no fixed spelling it is reliably retrievable by nobody; and it strips the verse of the hand in which it is conventionally written. The 1990 call for votes had listed literature and poetry as core subjects and languages as a subject of their own. The wire it was posted on could carry the discussion and not the script.

Who could post: the .pk delegation and what came after it

The country-code top-level domain .pk was registered in the IANA root on 3 June 1992, two years and three months after the group was created. Its sponsor and registry is PKNIC, a non-profit, non-statutory, member-based corporation established in June 1992 and the only organisation endorsed by the Government of Pakistan to administer .pk names. Registrations were made at the third level under second-level labels such as .com.pk, .edu.pk, .gov.pk and .net.pk; second-level names began to be opened up in 2005.

Internet access in Pakistan dates from the early 1990s, and the growth curve through the group's lifetime was slow by later standards. In 2001 about 1.3 per cent of the population used the Internet; by 2006 the figure was 6.5 per cent, and by 2012 10 per cent. For most of the years in which soc.culture.pakistan was busy, in other words, the overwhelming majority of people in Pakistan were not on the network at all, and the group's centre of gravity stayed abroad.

The reorganisation paperwork of 1995 shows the group's own participants expecting that to change, and quickly. The November 1995 proposal claimed traffic of over two hundred messages a day on the unmoderated group and predicted that in the next few months more Usenet readers would come online from Pakistan, estimated at more than a hundred thousand. Both figures are the proponent's own and neither is independently verifiable; they are recorded here as what the group's readers believed about themselves at the end of 1995, which is a different kind of fact and a useful one.

What is verifiable is what happened to the ballot rolls. The 1990 roll contained no .pk addresses because none existed. The final roll of the 1996 reorganisation vote, printed one line per voter in the same tradition, tells a different story: counting the addresses on it gives about 146 at .com hosts, about 145 at .edu, about 118 under .pk, some thirty in .uk and about nineteen in .ca, with Australia, Sweden, Japan, France, Finland, the Netherlands, Norway, Singapore, Ireland, Malaysia, Saudi Arabia and the United Arab Emirates all represented in single figures. Six years had produced a domestic readership from nothing.

The Pakistani hosts on that roll are worth naming, because they are the country's early commercial Internet in miniature. Of the 118 .pk addresses, 91 sit under erum.com.pk, and sixty of those under one host, biruni.erum.com.pk; the rest of the erum names carry city tags — khi, lhr, isb, psw — which is what a single provider with points of presence in Karachi, Lahore, Islamabad and Peshawar looks like on a voter list. Another twenty are spread across brain.com.pk and brain.net.pk, the domains of Brain Computer Services of Lahore and its Internet arm, Brain NET. Four are at lums.edu.pk, Lahore University of Management Sciences, and two at paknet1.ptc.pk, under the domain of the state telecommunications corporation. One further ballot came from a machine in the United Arab Emirates and one from a university in Saudi Arabia.

The diaspora, and where it was counted

The 1990 call for votes had listed "communities abroad, problems, needs" among the group's subjects, and for most of its life the group was, in practice, a diaspora institution: a way for people distributed across a dozen time zones to hold one continuous conversation about a country most of them were not standing in.

The distribution is well documented, though the figures are necessarily recent rather than contemporary. Estimates compiled by country put the overseas Pakistani population at roughly ten million as of 2022, of whom over 4.7 million are in the Middle East. Saudi Arabia recorded 1,814,678 Pakistani nationals in its 2022 census. The United Kingdom's 2021 census counted 1,662,286 people of Pakistani ethnicity, the second-largest community; the United Arab Emirates is estimated at about 1.6 million. The United States recorded 684,438 in the 2023 American Community Survey, concentrated around New York, Chicago and New Jersey. Kuwait, Canada, Oman and Qatar follow, and Italy, Germany, Spain, Norway and Australia all host substantial communities. The Ministry of Overseas Pakistanis put the figure at approximately 8.8 million in December 2017; data released in 2023 record more than 10.8 million Pakistanis moving abroad since 1990.

None of those numbers describes the group's readership, and none is offered as doing so. What they establish is the shape of the population the group drew on: heavily Gulf-based in absolute terms, heavily British and North American in the segment that had university network access in 1990. The ballot rolls are consistent with that. The Gulf communities appear on the 1996 roll in ones and twos; the North American universities appear in the hundreds. Usenet in the early 1990s was not a mirror of where Pakistanis lived. It was a mirror of where Pakistanis had free bandwidth.

The overseas dimension nearly got a group of its own. The October 1995 proposal included soc.culture.pakistan.overseas — "Overseas Pakistanis" — chartered for information and discussion among Pakistanis living abroad, and noting that the government maintained a special committee for them. Like the linguistics group, it was cut from the revised proposal and never voted on.

The 1995-96 reorganisation, and the vote that had to be run twice

By late 1995 the group was, on its own participants' account, too busy to be usable. On 23 October 1995 Asim Mughal posted a request for discussion proposing to reorganise soc.culture.pakistan by building a sub-hierarchy beneath it, leaving the original group untouched. The reasoning was stated without much diplomacy: traffic had increased steadily to the point where any meaningful discussion of any topic was lost in the volume and the noise, rendering the group unusable for serious readership and driving that readership away. The proposal was credited to discussions held in 1994 and 1995 by a Pakistan Networking Committee, and was cross-posted to news.announce.newgroups, news.groups, soc.culture.pakistan and bit.listserv.pns-l, the Usenet gateway of a mailing list whose name points to the Pakistan News Service; the same proposal credits that service with the software used to format the document.

The first proposal listed nine groups: education, history, info, linguistics, moderated, overseas, politics, religion and sports, three of them moderated. The second, posted 21 November 1995, cut it to six — education, history and moderated as moderated groups; politics, religion and sports unmoderated — and it is that six that went to a vote. Each charter is a small essay on what its authors thought the subject needed. The education charter argued that the state allocated less than one per cent of its budget to education and that private colleges and universities had mushroomed in consequence. The history charter dated independence from the British to 14 August 1947 and asked for an academic forum for the events that led to the creation of Pakistan, expressly not excluding anything since. The politics charter observed that the number of political parties exceeded ten and that constant coalitions and splits made political discussion very lively. The religion charter noted that Pakistan is predominantly Muslim, that citizens of other faiths hold reserved seats in parliament and have a separate federal ministry, and that calls for the enforcement of Islamic systems and laws had recently been made.

The vote, conducted by the Usenet Volunteer Votetakers, went badly wrong. A result posted on 15 February 1996 recorded all six groups passing, with margins around 230 to 250 in favour against 40 to 60 against — and it was posted with an unusually blunt note from the vote-taker, Jan Isley, recording that in his opinion serious irregularities had begun before the vote started and that considerable effort had been spent stuffing the ballot box. On 9 April 1996 David Lawrence, the moderator of news.announce.newgroups, declared the result invalid.

The invalidation notice is one of the more careful documents in the Usenet administrative record and deserves to be described accurately. Lawrence's stated reason was not that the proposal was bad but that the outcome could not be determined: with the problem ballots removed, every part of the proposal failed, a couple by as few as five votes, and since many or all of the groups might well have passed on properly cast votes, invalidating gave the proponents an immediate chance to pass them cleanly. The vote-taker's appended analysis recorded that all votes from one European institutional domain had been invalidated after evidence that they were probably machine-generated, and that one of the proposed moderators had multiple ballots invalidated for voting from several addresses under several names and for circulating edited ballots. Lawrence added, in terms, that the decision carried no bias regarding the content or the users of the proposed groups, that he had said more than once he was fine with the proposal passing, and that he had not voted against it.

The revote was run at once. Voting closed on 6 May 1996 and the result was posted on 24 May: all six groups passed, on much larger turnouts and much wider margins — education 479 to 69, history 470 to 84, moderated 467 to 75, politics 477 to 72, religion 465 to 81, sports 472 to 70, with fifteen invalid votes. After the mandatory five-day objection period the control messages went out, and the six groups were created on 30 May 1996. Counting from the first request for discussion, the reorganisation had taken about nine months, a figure the Big-8 board's own later summary repeats.

What the room carried, on the evidence of its own charters

The honest way to describe an unmoderated newsgroup whose archive is not preserved here is to describe what its own participants said it contained, in documents they wrote and signed, and then to stop. Fortunately those documents are unusually detailed, because six of them were written in 1995 by people trying to persuade several hundred of their fellow readers to vote for a new room.

Read together, the six charters and the 1990 subject list give a consistent picture of a general-interest national group with four heavy strands and a great deal else. The heavy strands were politics, religion, history and sport; the rest was language and literature, education, food, local news, events abroad, and the ordinary business of living between two countries. The call for votes in the 1996 revote added one further detail: it stated that the topics chosen for the new groups were the largest threads of discussion on the existing soc.culture.pakistan and on some Pakistani mailing lists, and that this was why those particular topics had been chosen. That sentence is as close to a contents page as the record gets, and it is the proponents' own account rather than an independent measurement.

Two of the new rooms tell us something about the group's role as a news relay. The Big-8 board's 2007 review of soc.culture.pakistan.moderated recorded that during its active period a large share of its content consisted of reposts from the Pakistan News Service rather than user-initiated discussion. In the years before Pakistani newspapers were reliably on the web, relaying news from home to readers with no other same-day source was among the most concrete services these groups performed — and it is the service the web most completely made redundant.

Cricket, hockey and squash

Cricket comes first because the group's own paperwork puts it there. The 1995 charter for soc.culture.pakistan.sports opens with the claim that Pakistan holds world championships in squash, hockey and cricket, much of which goes unnoticed and undiscussed, and asks for a room for commentary, critique, scores, rankings, upcoming events, equipment, coaches and clubs. The sports group passed 472 to 70.

The claim in that charter was accurate on all three counts. Pakistan won the Cricket World Cup on 25 March 1992, beating England by 22 runs in the final at the Melbourne Cricket Ground; Imran Khan, the captain, scored 72 in what was his last one-day international. In field hockey Pakistan holds a record four World Cup titles — 1971, 1978, 1982 and 1994 — and three Olympic gold medals, in 1960, 1968 and 1984. In squash Jahangir Khan won the World Open six times and the British Open ten years running from 1982 to 1991, and was unbeaten from 1981 to 1986 across 555 consecutive matches, the longest winning streak recorded in top-level professional sport; Jansher Khan won the World Open eight times and the British Open six, and held the world number one ranking from January 1988 to January 1998.

The timing is what made cricket a newsgroup subject rather than merely a national one. The 1992 World Cup win fell two years into the group's life. The 1996 World Cup was co-hosted by Pakistan, India and Sri Lanka, with the final played at the Gaddafi Stadium in Lahore on 17 March 1996 — ten weeks before the sports group was created. Pakistan reached the final of the 1999 World Cup at Lord's. Each of the three co-hosts of that tournament had a national group of its own in the same branch, and the fixture list put all three in the same weeks of the same years.

Poetry, music, film and the table

The 1990 subject list put literature, poetry, art and folklore third, ahead of science and technology, and the 1995 linguistics charter asked specifically for prose, poetry and ghazals, and for discussion of the works of poets and writers in Urdu, Sindhi, Punjabi, Pashto, Balochi and Saraiki. It is the strand of the subject on which the romanisation problem bore hardest, because in a tradition where the exact word and the exact metre are the point, the spelling is not incidental.

The ghazal is a form of amatory ode originating in Arabic poetry, concerned with love, loss and separation, and it is among the central forms of Urdu literary culture. Its social institution is the mushaira, a gathering at which poets recite publicly, often improvising and often competing; the mushaira has been a defining institution of Urdu literary culture in northern India, Pakistan and the Deccan, and is frequently described as a forum for free self-expression. A newsgroup is not a mushaira, and nothing in the surviving record says what form the group's literary threads took. What the record does establish is that a recital tradition and a text-only network make an awkward pair, and that the 1995 charter asked for a room in which prose, poetry and ghazals could be discussed.

The canon named in those charters is largely nineteenth and twentieth century. Mirza Ghalib wrote in Urdu and Persian in Delhi in the nineteenth century. Muhammad Iqbal, whose Urdu poetry is counted among the greatest of the twentieth century, is Pakistan's national poet, and his vision of a cultural and political ideal for the Muslims of British India is widely regarded as having animated the impulse behind the Pakistan movement. Faiz Ahmad Faiz, poet in Urdu and Punjabi, teacher, army officer, journalist, trade unionist and broadcaster, remains among the best known Urdu writers of his century in both Pakistan and India.

Qawwali, the Sufi devotional form performed at shrines across the subcontinent, had by the 1990s an international audience largely because of one performer: Nusrat Fateh Ali Khan, born 13 October 1948, credited with introducing qawwali to audiences outside South Asia, and described by the New York Times as the greatest qawwali singer of his generation. He died in London on 16 August 1997, at the centre of the years in which such groups were busiest.

Film had its own geography. Pakistani cinema is centred on Lahore, Karachi and Peshawar, and Lollywood, the Lahore-based industry that produces films in Urdu and Punjabi, is among the largest of its sub-industries. Food appears in the 1990 subject list in its own right — "food, cookery" — one of the nine headings on which the electorate of March 1990 was asked to vote.

The neighbours

soc.culture.pakistan sat in a South Asian cluster with soc.culture.indian and soc.culture.bangladesh. Both are older in the record than their sub-groups: the ISC news.announce.newgroups archive holds no creation paperwork for either parent group, only for the Indian group's many later offshoots, which is what one expects of groups that predate the archive's coverage. The 1990 request for discussion for the Pakistan group takes soc.culture.indian's existence entirely for granted and treats it as the room the traffic was currently coming out of.

The cluster mapped a partitioned region onto an unpartitioned network. Borders that were difficult to cross on the ground were crossed by a comma in a Newsgroups: header, and that was simultaneously the arrangement's chief value and its chief source of friction: the mechanism that could put readers from three successor states into a single thread also made any crisis in the region topical in all three rooms at once. The 2007 proposal to remove soc.culture.pakistan.history put the point in flat administrative language, noting that because of sometimes volatile relations with neighbouring states such as India and Bangladesh (the former East Pakistan), converting that dormant moderated group to unmoderated status might simply produce another group to include in a cross-posted thread disruptive to the others.

That pattern — the same argument conducted in several rooms at once, and the remedies the branch tried against it — is documented in this directory at soc.culture.kuwait.moderated, which owns that account and is not duplicated here. Bangladesh's own group, including the 1971 material seen from Dhaka, is at soc.culture.bangladesh.

How the branch ended

The sub-hierarchy did not last. Of the three moderated groups created in May 1996, the Big-8 Management Board's later review found that the last approved article specific to soc.culture.pakistan.education was posted in April 1998, the last specific to soc.culture.pakistan.history in July 1997, and the last specific to soc.culture.pakistan.moderated in June 1998. During their active periods the board estimated roughly twenty threads a month for education, fifty for history and forty for moderated. After the moderators stopped approving new discussion, each group carried about a dozen widely cross-posted approvals over the following year or two — for the history group the board gave the window as two years — and then only the periodic reposting of FAQs and guidelines, which the board recorded as continuing through 2006. That is roughly eight years of automated housekeeping over rooms in which no new discussion was being approved.

On 4 November 2001 an unauthorised rmgroup control message was sent for soc.culture.pakistan and for all six of its 1996 sub-groups by a third party who described them as bogus newsgroups. Messages of that kind carried no authority over Big-8 groups and were routinely ignored by administrators who checked signatures. They had no effect on the namespace.

The end of the moderated groups was procedurally correct and unhurried. On 17 April 2007 Jim Riley posted three separate requests for discussion, one per group, expressly noting that the three need not share the same fate: one might be removed, one converted to unmoderated status, one given a new moderator. Each RFD set out the group's traffic history, reprinted its charter from the 1996 call for votes, and listed the alternatives. Last calls for comments went out on 1 May; after the five-day comment period the Big-8 Management Board decided by consensus, and results were posted on 8 May 2007. All three groups were removed on 11 May 2007, by rmgroup control messages issued from [email protected]. Removal notices reissued through May and June, as such messages routinely were, to catch servers that had missed the first.

What survives in the namespace today is the parent group and half of its 1996 offspring. The current ISC newsgroups file lists soc.culture.pakistan, soc.culture.pakistan.politics, soc.culture.pakistan.religion and soc.culture.pakistan.sports, and no others. The three groups that were removed are the three that had been created as moderated.

What removed the three moderated rooms was procedure; what emptied them was the same drift that emptied most of the hierarchy. From the late 1990s web forums and mailing lists drew off the practical traffic, unmoderated groups accumulated automated advertising faster than conversation, and the social platforms of the 2000s absorbed much of what remained. soc.culture.pakistan itself was never formally closed — Usenet groups rarely are — which is why it and its three unmoderated sub-groups are still marked postable. Whether anything is being posted to them is not something the newsgroups and active files record, and this page does not guess.

What the record does not show

It is worth being explicit about the size of the gaps, because they are large and because the temptation to fill them with plausible colour is exactly what would make this page worthless.

There is no readership figure for soc.culture.pakistan. The Arbitron-style Usenet readership surveys of the early 1990s covered many groups, but no figure for this one is quoted in any source consulted for this article, and none is invented here. The claim of more than two hundred messages a day comes from a proposal document written by an interested party in 1995 and is reported as such.

There is no record here of who the group's regular posters were, what its FAQ said, who maintained it, or how many versions it went through. The FAQ's existence is documented — the Big-8 board's 2007 review records FAQs and guidelines being posted into the sub-groups as late as 2006 — but its contents are not in the administrative archive and are not reproduced or characterised.

The names that do appear on this page appear because their bearers put them on public administrative postings: the proposer, the vote-takers, the moderator of news.announce.newgroups, the proponents of the reorganisation and of the removals. The 1996 invalidation notice names an individual in connection with ballot irregularities; that allegation is a matter of public record in the archive, and this page describes the finding without repeating the name, because the finding is what bears on the group's history and the name is not.

Nothing here should be read as an account of what any individual posted. The group ran unmoderated throughout, across a period in which almost every question it touched was actively disputed; a page of this kind can say what the room was for and what its own paperwork said it contained, and it should not pretend to say more.

Scope, sources and limits

No part of this article is based on the group's own traffic. Nothing in it quotes, counts, characterises or summarises any archived posting, and no such posting has been read in preparing it. This directory holds no other page for soc.culture.pakistan, and there is no companion page carrying the messages elsewhere on this site.

Everything quoted above comes from two administrative sources that are separate from the group's traffic and are publicly mirrored: the news.announce.newgroups archive, which holds the 1990 creation file, the 1995-96 reorganisation file and the three 2007 removal files for this group, and the Usenet control-message archive, which holds the newgroup and rmgroup messages and their dates. The current namespace comes from the newsgroups and active files in the same archive. Country-code delegation is from the IANA root database. The remaining factual matter — dates, census proportions, encoding versions, sporting results — is drawn from general reference works and is cited in the text where it carries weight.

Two limits of interpretation apply throughout. First, the ballot rolls counted in this article are rolls of people who voted, not of people who read; the domain proportions given for 1990 and 1996 are counts of addresses printed on those rolls, made for this article by parsing the rolls themselves, and are offered as evidence of where network access existed rather than as measurements of readership. The 1990 count is complete; the 1996 count reaches 522 of the 557 valid ballots the vote-taker reported, so its figures are given as approximations. Second, on every contested question this page names — partition, 1971, Kashmir, the role of the army, the place of religion in the state, the language question — the intention is to record that these were argued in the group and to give the dates around which they were argued. No position among them is endorsed, ranked or resolved here, and none should be inferred from the order in which they appear.